4 applications ofdifferentiationcollege.cengage.com/mathematics/1111007888_stewart... · 4...
TRANSCRIPT
4 APPLICATIONS OF DIFFERENTIATION4.1 Maximum and Minimum Values
1. A function f has an absolute minimum at x = c if f(c) is the smallest function value on the entire domain of f , whereas
f has a local minimum at c if f(c) is the smallest function value when x is near c.
3. Absolute maximum at s, absolute minimum at r, local maximum at c, local minima at b and r, neither a maximum nor a
minimum at a and d.
5. Absolute maximum value is f(4) = 5; there is no absolute minimum value; local maximum values are f(4) = 5 and
f(6) = 4; local minimum values are f(2) = 2 and f(1) = f(5) = 3.
7. Absolute minimum at 2, absolute maximum at 3,
local minimum at 4
9. Absolute maximum at 5, absolute minimum at 2,
local maximum at 3, local minima at 2 and 4
11. (a) (b) (c)
13. (a) Note: By the Extreme Value Theorem,
f must not be continuous; because if it
were, it would attain an absolute
minimum.
(b)
111
112 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
15. f(x) = 8− 3x, x ≥ 1. Absolute maximum f(1) = 5; no
local maximum. No absolute or local minimum.
17. f(x) = x2, 0 < x < 2. No absolute or local maximum or
minimum value.
19. f(x) = x2, 0 ≤ x < 2. Absolute minimum
f(0) = 0; no local minimum. No absolute or
local maximum.
21. f(x) = x2, −3 ≤ x ≤ 2. Absolute maximum
f(−3) = 9. No local maximum. Absolute and
local minimum f(0) = 0.
23. f(t) = 1/t, 0 < t < 1. No maximum or minimum. 25. f(x) = 1−√x. Absolute maximum f(0) = 1; no local
maximum. No absolute or local minimum.
27. f(x) =1− x if 0 ≤ x < 2
2x− 4 if 2 ≤ x ≤ 3
Absolute maximum f(3) = 2; no local maximum. No absolute or local minimum.
SECTION 4.1 MAXIMUM AND MINIMUM VALUES ¤ 113
29. f(x) = 5x2 + 4x ⇒ f 0(x) = 10x+ 4. f 0(x) = 0 ⇒ x = − 25 , so − 2
5 is the only critical number.
31. f(x) = x3 + 3x2 − 24x ⇒ f 0(x) = 3x2 + 6x− 24 = 3(x2 + 2x− 8).f 0(x) = 0 ⇒ 3(x+ 4)(x− 2) = 0 ⇒ x = −4, 2. These are the only critical numbers.
33. s(t) = 3t4 + 4t3 − 6t2 ⇒ s0(t) = 12t3 + 12t2 − 12t. s0(t) = 0 ⇒ 12t(t2 + t− 1) ⇒
t = 0 or t2 + t− 1 = 0. Using the quadratic formula to solve the latter equation gives us
t =−1± 12 − 4(1)(−1)
2(1)=−1±√5
2≈ 0.618, −1.618. The three critical numbers are 0, −1±
√5
2.
35. g(y) = y − 1y2 − y + 1
⇒
g0(y) =(y2 − y + 1)(1)− (y − 1)(2y − 1)
(y2 − y + 1)2=
y2 − y + 1− (2y2 − 3y + 1)(y2 − y + 1)2
=−y2 + 2y
(y2 − y + 1)2=
y(2− y)
(y2 − y + 1)2.
g0(y) = 0 ⇒ y = 0, 2. The expression y2 − y + 1 is never equal to 0, so g0(y) exists for all real numbers.
The critical numbers are 0 and 2.
37. h(t) = t3/4 − 2t1/4 ⇒ h0(t) = 34t−1/4 − 2
4t−3/4 = 1
4t−3/4(3t1/2 − 2) = 3
√t− 2
44√t3
.
h0(t) = 0 ⇒ 3√t = 2 ⇒ √
t = 23⇒ t = 4
9. h0(t) does not exist at t = 0, so the critical numbers are 0 and 4
9.
39. F (x) = x4/5(x− 4)2 ⇒
F 0(x) = x4/5 · 2(x− 4) + (x− 4)2 · 45x−1/5 = 1
5x−1/5(x− 4)[5 · x · 2 + (x− 4) · 4]
=(x− 4)(14x− 16)
5x1/5=2(x− 4)(7x− 8)
5x1/5
F 0(x) = 0 ⇒ x = 4, 87
. F 0(0) does not exist. Thus, the three critical numbers are 0, 87
, and 4.
41. f(θ) = 2 cos θ + sin2 θ ⇒ f 0(θ) = −2 sin θ + 2 sin θ cos θ. f 0(θ) = 0 ⇒ 2 sin θ (cos θ − 1) = 0 ⇒ sin θ = 0
or cos θ = 1 ⇒ θ = nπ [n an integer] or θ = 2nπ. The solutions θ = nπ include the solutions θ = 2nπ, so the critical
numbers are θ = nπ.
43. A graph of f 0(x) = 1 + 210 sinx
x2 − 6x+ 10 is shown. There are 10 zeros
between −25 and 25 (one is approximately −0.05). f 0 exists
everywhere, so f has 10 critical numbers.
45. f(x) = 3x2 − 12x+ 5, [0, 3]. f 0(x) = 6x− 12 = 0 ⇔ x = 2. Applying the Closed Interval Method, we find that
f(0) = 5, f(2) = −7, and f(3) = −4. So f(0) = 5 is the absolute maximum value and f(2) = −7 is the absolute minimum
value.
114 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
47. f(x) = 2x3 − 3x2 − 12x+ 1, [−2, 3]. f 0(x) = 6x2 − 6x− 12 = 6(x2 − x− 2) = 6(x− 2)(x+ 1) = 0 ⇔x = 2,−1. f(−2) = −3, f(−1) = 8, f(2) = −19, and f(3) = −8. So f(−1) = 8 is the absolute maximum value and
f(2) = −19 is the absolute minimum value.
49. f(x) = x4 − 2x2 + 3, [−2, 3]. f 0(x) = 4x3 − 4x = 4x(x2 − 1) = 4x(x+ 1)(x− 1) = 0 ⇔ x = −1, 0, 1.
f(−2) = 11, f(−1) = 2, f(0) = 3, f(1) = 2, f(3) = 66. So f(3) = 66 is the absolute maximum value and f(±1) = 2 is
the absolute minimum value.
51. f(x) = x
x2 + 1, [0, 2]. f 0(x) =
(x2 + 1)− x(2x)
(x2 + 1)2=
1− x2
(x2 + 1)2= 0 ⇔ x = ±1, but −1 is not in [0, 2]. f(0) = 0,
f(1) = 12
, f(2) = 25
. So f(1) = 12
is the absolute maximum value and f(0) = 0 is the absolute minimum value.
53. f(t) = t√4− t2, [−1, 2].
f 0(t) = t · 12(4− t2)−1/2 (−2t) + (4− t2)1/2 · 1 = −t2√
4− t2+√4− t2 =
−t2 + (4− t2)√4− t2
=4− 2t2√4− t2
.
f 0(t) = 0 ⇒ 4− 2t2 = 0 ⇒ t2 = 2 ⇒ t = ±√2, but t = −√2 is not in the given interval, [−1, 2].f 0(t) does not exist if 4− t2 = 0 ⇒ t = ±2, but −2 is not in the given interval. f(−1) = −√3, f
√2 = 2, and
f(2) = 0. So f√2 = 2 is the absolute maximum value and f(−1) = −√3 is the absolute minimum value.
55. f(t) = 2 cos t+ sin 2t, [0, π/2].
f 0(t) = −2 sin t+ cos 2t · 2 = −2 sin t+ 2(1− 2 sin2 t) = −2(2 sin2 t+ sin t− 1) = −2(2 sin t− 1)(sin t+ 1).f 0(t) = 0 ⇒ sin t = 1
2or sin t = −1 ⇒ t = π
6. f(0) = 2, f(π
6) =
√3 + 1
2
√3 = 3
2
√3 ≈ 2.60, and f(π
2) = 0.
So f(π6) = 3
2
√3 is the absolute maximum value and f(π
2) = 0 is the absolute minimum value.
57. f(x) = xa(1− x)b, 0 ≤ x ≤ 1, a > 0, b > 0.
f 0(x) = xa · b(1− x)b−1(−1) + (1− x)b · axa−1 = xa−1(1− x)b−1[x · b(−1) + (1− x) · a]= xa−1(1− x)b−1(a− ax− bx)
At the endpoints, we have f(0) = f(1) = 0 [the minimum value of f ]. In the interval (0, 1), f 0(x) = 0 ⇔ x =a
a+ b.
fa
a+ b=
a
a+ b
a
1− a
a+ b
b
=aa
(a+ b)aa+ b− a
a+ b
b
=aa
(a+ b)a· bb
(a+ b)b=
aabb
(a+ b)a+b.
So f a
a+ b=
aabb
(a+ b)a+bis the absolute maximum value.
59. (a) From the graph, it appears that the absolute maximum value is about
f(−0.77) = 2.19, and the absolute minimum value is about
f(0.77) = 1.81.
SECTION 4.1 MAXIMUM AND MINIMUM VALUES ¤ 115
(b) f(x) = x5 − x3 + 2 ⇒ f 0(x) = 5x4 − 3x2 = x2(5x2 − 3). So f 0(x) = 0 ⇒ x = 0, ± 35 .
f − 35
= − 35
5
− − 35
3
+ 2 = − 35
2 35 +
35
35 + 2 =
35 − 9
2535 + 2 =
625
35 + 2 (maximum)
and similarly, f 35
= − 625
35+ 2 (minimum).
61. (a) From the graph, it appears that the absolute maximum value is about
f(0.75) = 0.32, and the absolute minimum value is f(0) = f(1) = 0;
that is, at both endpoints.
(b) f(x) = x√x− x2 ⇒ f 0(x) = x · 1− 2x
2√x− x2
+√x− x2 =
(x− 2x2) + (2x− 2x2)2√x− x2
=3x− 4x22√x− x2
.
So f 0(x) = 0 ⇒ 3x− 4x2 = 0 ⇒ x(3− 4x) = 0 ⇒ x = 0 or 34
.
f(0) = f(1) = 0 (minimum), and f 34= 3
434− 3
4
2= 3
4316= 3
√3
16(maximum).
63. The density is defined as ρ = massvolume
=1000
V (T )(in g/cm3). But a critical point of ρ will also be a critical point of V
[since dρ
dT= −1000V −2 dV
dTand V is never 0], and V is easier to differentiate than ρ.
V (T ) = 999.87− 0.06426T + 0.0085043T 2 − 0.0000679T 3 ⇒ V 0(T ) = −0.06426 + 0.0170086T − 0.0002037T 2.
Setting this equal to 0 and using the quadratic formula to find T , we get
T =−0.0170086±√0.01700862 − 4 · 0.0002037 · 0.06426
2(−0.0002037) ≈ 3.9665◦C or 79.5318◦C. Since we are only interested
in the region 0◦C ≤ T ≤ 30◦C, we check the density ρ at the endpoints and at 3.9665◦C: ρ(0) ≈ 1000
999.87≈ 1.00013;
ρ(30) ≈ 1000
1003.7628≈ 0.99625; ρ(3.9665) ≈ 1000
999.7447≈ 1.000255. So water has its maximum density at
about 3.9665◦C.
65. Let a = −0.000 032 37, b = 0.000 903 7, c = −0.008 956, d = 0.03629, e = −0.04458, and f = 0.4074.
Then S(t) = at5 + bt4 + ct3 + dt2 + et+ f and S0(t) = 5at4 + 4bt3 + 3ct2 + 2dt+ e.
We now apply the Closed Interval Method to the continuous function S on the interval 0 ≤ t ≤ 10. Since S0 exists for all t,
the only critical numbers of S occur when S0(t) = 0. We use a rootfinder on a CAS (or a graphing device) to find that
S0(t) = 0 when t1 ≈ 0.855, t2 ≈ 4.618, t3 ≈ 7.292, and t4 ≈ 9.570. The values of S at these critical numbers are
S(t1) ≈ 0.39, S(t2) ≈ 0.43645, S(t3) ≈ 0.427, and S(t4) ≈ 0.43641. The values of S at the endpoints of the interval are
S(0) ≈ 0.41 and S(10) ≈ 0.435. Comparing the six numbers, we see that sugar was most expensive at t2 ≈ 4.618
(corresponding roughly to March 1998) and cheapest at t1 ≈ 0.855 (June 1994).
116 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
67. (a) v(r) = k(r0 − r)r2 = kr0r2 − kr3 ⇒ v0(r) = 2kr0r − 3kr2. v0(r) = 0 ⇒ kr(2r0 − 3r) = 0 ⇒
r = 0 or 23r0 (but 0 is not in the interval). Evaluating v at 1
2r0, 2
3r0, and r0, we get v 1
2r0 = 1
8kr30 , v 2
3r0 = 4
27kr30 ,
and v(r0) = 0. Since 427 >
18 , v attains its maximum value at r = 2
3r0. This supports the statement in the text.
(b) From part (a), the maximum value of v is 427kr30 .
(c)
69. f(x) = x101 + x51 + x+ 1 ⇒ f 0(x) = 101x100 + 51x50 + 1 ≥ 1 for all x, so f 0(x) = 0 has no solution. Thus, f(x)
has no critical number, so f(x) can have no local maximum or minimum.
71. If f has a local minimum at c, then g(x) = −f(x) has a local maximum at c, so g0(c) = 0 by the case of Fermat’s Theorem
proved in the text. Thus, f 0(c) = −g0(c) = 0.
4.2 The Mean Value Theorem
1. f(x) = 5− 12x+ 3x2, [1, 3]. Since f is a polynomial, it is continuous and differentiable on R, so it is continuous on [1, 3]
and differentiable on (1, 3). Also f(1) = −4 = f(3). f 0(c) = 0 ⇔ −12 + 6c = 0 ⇔ c = 2, which is in the open
interval (1, 3), so c = 2 satisfies the conclusion of Rolle’s Theorem.
3. f(x) =√x− 1
3x, [0, 9]. f , being the difference of a root function and a polynomial, is continuous and differentiable
on [0,∞), so it is continuous on [0, 9] and differentiable on (0, 9). Also, f(0) = 0 = f(9). f 0(c) = 0 ⇔1
2√c− 1
3= 0 ⇔ 2
√c = 3 ⇔ √
c =3
2⇒ c =
9
4, which is in the open interval (0, 9), so c = 9
4satisfies the
conclusion of Rolle’s Theorem.
5. f(x) = 1− x2/3. f(−1) = 1− (−1)2/3 = 1− 1 = 0 = f(1). f 0(x) = − 23x−1/3, so f 0(c) = 0 has no solution. This
does not contradict Rolle’s Theorem, since f 0(0) does not exist, and so f is not differentiable on (−1, 1).
7. f(8)− f(0)
8− 0 =6− 48
=1
4. The values of c which satisfy f 0(c) = 1
4seem to be about c = 0.8, 3.2, 4.4, and 6.1.
SECTION 4.2 THE MEAN VALUE THEOREM ¤ 117
9. (a), (b) The equation of the secant line is
y − 5 = 8.5− 58− 1 (x− 1) ⇔ y = 1
2x+92 .
(c) f(x) = x+ 4/x ⇒ f 0(x) = 1− 4/x2.
So f 0(c) = 12⇒ c2 = 8 ⇒ c = 2
√2, and
f(c) = 2√2 + 4
2√2= 3
√2. Thus, an equation of the
tangent line is y − 3√2 = 12x− 2√2 ⇔
y = 12x+ 2
√2.
11. f(x) = 3x2 + 2x+ 5, [−1, 1]. f is continuous on [−1, 1] and differentiable on (−1, 1) since polynomials are continuous
and differentiable on R. f 0(c) =f(b)− f(a)
b− a⇔ 6c+ 2 =
f(1)− f(−1)1− (−1) =
10− 62
= 2 ⇔ 6c = 0 ⇔
c = 0, which is in (−1, 1).
13. f(x) = 3√x, [0, 1]. f is continuous on R and differentiable on (−∞, 0) ∪ (0,∞), so f is continuous on [0, 1]
and differentiable on (0, 1). f 0(c) =f(b)− f(a)
b− a⇔ 1
3c2/3=
f(1)− f(0)
1− 0 ⇔ 1
3c2/3=1− 01
⇔
3c2/3 = 1 ⇔ c2/3 = 13⇔ c2 = 1
3
3= 1
27⇔ c = ± 1
27= ±
√39
, but only√39
is in (0, 1).
15. f(x) = (x− 3)−2 ⇒ f 0(x) = −2 (x− 3)−3. f(4)− f(1) = f 0(c)(4− 1) ⇒ 1
12− 1
(−2)2 =−2
(c− 3)3 · 3 ⇒
3
4=
−6(c− 3)3 ⇒ (c− 3)3 = −8 ⇒ c− 3 = −2 ⇒ c = 1, which is not in the open interval (1, 4). This does not
contradict the Mean Value Theorem since f is not continuous at x = 3.
17. Let f(x) = 1 + 2x+ x3 + 4x5. Then f(−1) = −6 < 0 and f(0) = 1 > 0. Since f is a polynomial, it is continuous, so the
Intermediate Value Theorem says that there is a number c between −1 and 0 such that f(c) = 0. Thus, the given equation has
a real root. Suppose the equation has distinct real roots a and b with a < b. Then f(a) = f(b) = 0. Since f is a polynomial, it
is differentiable on (a, b) and continuous on [a, b]. By Rolle’s Theorem, there is a number r in (a, b) such that f 0(r) = 0. But
f 0(x) = 2 + 3x2 + 20x4 ≥ 2 for all x, so f 0(x) can never be 0. This contradiction shows that the equation can’t have two
distinct real roots. Hence, it has exactly one real root.
19. Let f(x) = x3 − 15x+ c for x in [−2, 2]. If f has two real roots a and b in [−2, 2], with a < b, then f(a) = f(b) = 0. Since
the polynomial f is continuous on [a, b] and differentiable on (a, b), Rolle’s Theorem implies that there is a number r in (a, b)
118 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
such that f 0(r) = 0. Now f 0(r) = 3r2 − 15. Since r is in (a, b), which is contained in [−2, 2], we have |r| < 2, so r2 < 4.
It follows that 3r2 − 15 < 3 · 4− 15 = −3 < 0. This contradicts f 0(r) = 0, so the given equation can’t have two real roots
in [−2, 2]. Hence, it has at most one real root in [−2, 2].
21. (a) Suppose that a cubic polynomial P (x) has roots a1 < a2 < a3 < a4, so P (a1) = P (a2) = P (a3) = P (a4).
By Rolle’s Theorem there are numbers c1, c2, c3 with a1 < c1 < a2, a2 < c2 < a3 and a3 < c3 < a4 and
P 0(c1) = P 0(c2) = P 0(c3) = 0. Thus, the second-degree polynomial P 0(x) has three distinct real roots, which is
impossible.
(b) We prove by induction that a polynomial of degree n has at most n real roots. This is certainly true for n = 1. Suppose
that the result is true for all polynomials of degree n and let P (x) be a polynomial of degree n+ 1. Suppose that P (x) has
more than n+ 1 real roots, say a1 < a2 < a3 < · · · < an+1 < an+2. Then P (a1) = P (a2) = · · · = P (an+2) = 0.
By Rolle’s Theorem there are real numbers c1, . . . , cn+1 with a1 < c1 < a2, . . . , an+1 < cn+1 < an+2 and
P 0(c1) = · · · = P 0(cn+1) = 0. Thus, the nth degree polynomial P 0(x) has at least n+ 1 roots. This contradiction shows
that P (x) has at most n+ 1 real roots.
23. By the Mean Value Theorem, f(4)− f(1) = f 0(c)(4− 1) for some c ∈ (1, 4). But for every c ∈ (1, 4) we have
f 0(c) ≥ 2. Putting f 0(c) ≥ 2 into the above equation and substituting f(1) = 10, we get
f(4) = f(1) + f 0(c)(4− 1) = 10 + 3f 0(c) ≥ 10 + 3 · 2 = 16. So the smallest possible value of f(4) is 16.
25. Suppose that such a function f exists. By the Mean Value Theorem there is a number 0 < c < 2 with
f 0(c) =f(2)− f(0)
2− 0 =5
2. But this is impossible since f 0(x) ≤ 2 < 5
2for all x, so no such function can exist.
27. We use Exercise 26 with f(x) =√1 + x, g(x) = 1 + 1
2x, and a = 0. Notice that f(0) = 1 = g(0) and
f 0(x) =1
2√1 + x
<1
2= g0(x) for x > 0. So by Exercise 26, f(b) < g(b) ⇒ √
1 + b < 1 + 12b for b > 0.
Another method: Apply the Mean Value Theorem directly to either f(x) = 1 + 12x−√1 + x or g(x) =
√1 + x on [0, b].
29. Let f(x) = sinx and let b < a. Then f(x) is continuous on [b, a] and differentiable on (b, a). By the Mean Value Theorem,
there is a number c ∈ (b, a) with sin a− sin b = f(a)− f(b) = f 0(c)(a− b) = (cos c)(a− b). Thus,
|sin a− sin b| ≤ |cos c| |b− a| ≤ |a− b|. If a < b, then |sin a− sin b| = |sin b− sina| ≤ |b− a| = |a− b|. If a = b, both
sides of the inequality are 0.
31. For x > 0, f(x) = g(x), so f 0(x) = g0(x). For x < 0, f 0(x) = (1/x)0 = −1/x2 and g0(x) = (1 + 1/x)0 = −1/x2, so
again f 0(x) = g0(x). However, the domain of g(x) is not an interval [it is (−∞, 0) ∪ (0,∞)] so we cannot conclude that
f − g is constant (in fact it is not).
SECTION 4.3 HOW DERIVATIVES AFFECT THE SHAPE OF A GRAPH ¤ 119
33. Let g(t) and h(t) be the position functions of the two runners and let f(t) = g(t)− h(t). By hypothesis,
f(0) = g(0)− h(0) = 0 and f(b) = g(b)− h(b) = 0, where b is the finishing time. Then by the Mean Value Theorem,
there is a time c, with 0 < c < b, such that f 0(c) = f(b)− f(0)
b− 0 . But f(b) = f(0) = 0, so f 0(c) = 0. Since
f 0(c) = g0(c)− h0(c) = 0, we have g0(c) = h0(c). So at time c, both runners have the same speed g0(c) = h0(c).
4.3 How Derivatives Affect the Shape of a Graph
1. (a) f is increasing on (1, 3) and (4, 6). (b) f is decreasing on (0, 1) and (3, 4).
(c) f is concave upward on (0, 2). (d) f is concave downward on (2, 4) and (4, 6).
(e) The point of inflection is (2, 3).
3. (a) Use the Increasing/Decreasing (I/D) Test. (b) Use the Concavity Test.
(c) At any value of x where the concavity changes, we have an inflection point at (x, f(x)).
5. (a) Since f 0(x) > 0 on (1, 5), f is increasing on this interval. Since f 0(x) < 0 on (0, 1) and (5, 6), f is decreasing on these
intervals.
(b) Since f 0(x) = 0 at x = 1 and f 0 changes from negative to positive there, f changes from decreasing to increasing and has
a local minimum at x = 1. Since f 0(x) = 0 at x = 5 and f 0 changes from positive to negative there, f changes from
increasing to decreasing and has a local maximum at x = 5.
7. There is an inflection point at x = 1 because f 00(x) changes from negative to positive there, and so the graph of f changes
from concave downward to concave upward. There is an inflection point at x = 7 because f 00(x) changes from positive to
negative there, and so the graph of f changes from concave upward to concave downward.
9. (a) f(x) = 2x3 + 3x2 − 36x ⇒ f 0(x) = 6x2 + 6x− 36 = 6(x2 + x− 6) = 6(x+ 3)(x− 2).We don’t need to include the “6” in the chart to determine the sign of f 0(x).
Interval x+ 3 x− 2 f 0(x) f
x < −3 − − + increasing on (−∞,−3)−3 < x < 2 + − − decreasing on (−3, 2)x > 2 + + + increasing on (2,∞)
(b) f changes from increasing to decreasing at x = −3 and from decreasing to increasing at x = 2. Thus, f(−3) = 81 is a
local maximum value and f(2) = −44 is a local minimum value.
(c) f 0(x) = 6x2 + 6x− 36 ⇒ f 00(x) = 12x+ 6. f 00(x) = 0 at x = − 12
, f 00(x) > 0 ⇔ x > − 12
, and
f 00(x) < 0 ⇔ x < − 12 . Thus, f is concave upward on − 1
2 ,∞ and concave downward on −∞,− 12
. There is an
inflection point at − 12, f − 1
2= − 1
2, 372
.
120 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
11. (a) f(x) = x4 − 2x2 + 3 ⇒ f 0(x) = 4x3 − 4x = 4x x2 − 1 = 4x(x+ 1)(x− 1).
Interval x+ 1 x x− 1 f 0(x) f
x < −1 − − − − decreasing on (−∞,−1)−1 < x < 0 + − − + increasing on (−1, 0)0 < x < 1 + + − − decreasing on (0, 1)
x > 1 + + + + increasing on (1,∞)
(b) f changes from increasing to decreasing at x = 0 and from decreasing to increasing at x = −1 and x = 1. Thus,
f(0) = 3 is a local maximum value and f(±1) = 2 are local minimum values.
(c) f 00(x) = 12x2 − 4 = 12 x2 − 13= 12 x+ 1/
√3 x− 1/√3 . f 00(x) > 0 ⇔ x < −1/√3 or x > 1/
√3 and
f 00(x) < 0 ⇔ −1/√3 < x < 1/√3. Thus, f is concave upward on −∞,−√3/3 and
√3/3,∞ and concave
downward on −√3/3,√3/3 . There are inflection points at ±√3/3, 229
.
13. (a) f(x) = sinx+ cosx, 0 ≤ x ≤ 2π. f 0(x) = cosx− sinx = 0 ⇒ cosx = sinx ⇒ 1 =sinx
cosx⇒
tanx = 1 ⇒ x = π4
or 5π4
. Thus, f 0(x) > 0 ⇔ cosx− sinx > 0 ⇔ cosx > sinx ⇔ 0 < x < π4
or
5π4< x < 2π and f 0(x) < 0 ⇔ cosx < sinx ⇔ π
4< x < 5π
4. So f is increasing on 0, π
4and 5π
4, 2π and f
is decreasing on π4, 5π4
.
(b) f changes from increasing to decreasing at x = π4
and from decreasing to increasing at x = 5π4
. Thus, f π4=√2 is a
local maximum value and f 5π4
= −√2 is a local minimum value.
(c) f 00(x) = − sinx− cosx = 0 ⇒ − sinx = cosx ⇒ tanx = −1 ⇒ x = 3π4
or 7π4
. Divide the interval
(0, 2π) into subintervals with these numbers as endpoints and complete a second derivative chart.
Interval f 00(x) = − sinx− cosx Concavity
0, 3π4
f 00 π2= −1 < 0 downward
3π4, 7π4
f 00(π) = 1 > 0 upward7π4, 2π f 00 11π
6= 1
2− 1
2
√3 < 0 downward
There are inflection points at 3π4, 0 and 7π
4, 0 .
15. f(x) = x5 − 5x+ 3 ⇒ f 0(x) = 5x4 − 5 = 5(x2 + 1)(x+ 1)(x− 1).First Derivative Test: f 0(x) < 0 ⇒ −1 < x < 1 and f 0(x) > 0 ⇒ x > 1 or x < −1. Since f 0 changes from
positive to negative at x = −1, f(−1) = 7 is a local maximum value; and since f 0 changes from negative to positive at x = 1,
f(1) = −1 is a local minimum value.
Second Derivative Test: f 00(x) = 20x3. f 0(x) = 0 ⇔ x = ±1. f 00(−1) = −20 < 0 ⇒ f(−1) = 7 is a local
maximum value. f 00(1) = 20 > 0 ⇒ f(1) = −1 is a local minimum value.
Preference: For this function, the two tests are equally easy.
SECTION 4.3 HOW DERIVATIVES AFFECT THE SHAPE OF A GRAPH ¤ 121
17. f(x) = x+√1− x ⇒ f 0(x) = 1 + 1
2 (1− x)−1/2(−1) = 1− 1
2√1− x
. Note that f is defined for 1− x ≥ 0; that is,
for x ≤ 1. f 0(x) = 0 ⇒ 2√1− x = 1 ⇒ √
1− x = 12⇒ 1− x = 1
4⇒ x = 3
4. f 0 does not exist at x = 1,
but we can’t have a local maximum or minimum at an endpoint.
First Derivative Test: f 0(x) > 0 ⇒ x < 34
and f 0(x) < 0 ⇒ 34< x < 1. Since f 0 changes from positive to
negative at x = 34
, f 34= 5
4is a local maximum value.
Second Derivative Test: f 00(x) = − 12− 12(1− x)−3/2(−1) = − 1
4√1− x
3 .
f 00 34= −2 < 0 ⇒ f 3
4= 5
4 is a local maximum value.
Preference: The First Derivative Test may be slightly easier to apply in this case.
19. (a) By the Second Derivative Test, if f 0(2) = 0 and f 00(2) = −5 < 0, f has a local maximum at x = 2.
(b) If f 0(6) = 0, we know that f has a horizontal tangent at x = 6. Knowing that f 00(6) = 0 does not provide any additional
information since the Second Derivative Test fails. For example, the first and second derivatives of y = (x− 6)4,
y = −(x− 6)4, and y = (x− 6)3 all equal zero for x = 6, but the first has a local minimum at x = 6, the second has a
local maximum at x = 6, and the third has an inflection point at x = 6.
21. f 0(0) = f 0(2) = f 0(4) = 0 ⇒ horizontal tangents at x = 0, 2, 4.
f 0(x) > 0 if x < 0 or 2 < x < 4 ⇒ f is increasing on (−∞, 0) and (2, 4).
f 0(x) < 0 if 0 < x < 2 or x > 4 ⇒ f is decreasing on (0, 2) and (4,∞).f 00(x) > 0 if 1 < x < 3 ⇒ f is concave upward on (1, 3).
f 00(x) < 0 if x < 1 or x > 3 ⇒ f is concave downward on (−∞, 1) and (3,∞).There are inflection points when x = 1 and 3.
23. f 0(x) > 0 if |x| < 2 ⇒ f is increasing on (−2, 2).f 0(x) < 0 if |x| > 2 ⇒ f is decreasing on (−∞,−2)and (2,∞). f 0(−2) = 0 ⇒ horizontal tangent at x = −2.
limx→2
|f 0(x)| =∞ ⇒ there is a vertical asymptote or
vertical tangent (cusp) at x = 2. f 00(x) > 0 if x 6= 2 ⇒f is concave upward on (−∞, 2) and (2,∞).
25. The function must be always decreasing (since the first derivative is always negative)
and concave downward (since the second derivative is always negative).
122 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
27. (a) f is increasing where f 0 is positive, that is, on (0, 2), (4, 6), and (8,∞); and decreasing where f 0 is negative, that is,
on (2, 4) and (6, 8).
(b) f has local maxima where f 0 changes from positive to negative, at x = 2 and at
x = 6, and local minima where f 0 changes from negative to positive, at x = 4
and at x = 8.
(e)
(c) f is concave upward (CU) where f 0 is increasing, that is, on (3, 6) and (6,∞),and concave downward (CD) where f 0 is decreasing, that is, on (0, 3).
(d) There is a point of inflection where f changes from being CD to being CU, that
is, at x = 3.
29. (a) f(x) = 2x3 − 3x2 − 12x ⇒ f 0(x) = 6x2 − 6x− 12 = 6(x2 − x− 2) = 6(x− 2)(x+ 1).f 0(x) > 0 ⇔ x < −1 or x > 2 and f 0(x) < 0 ⇔ −1 < x < 2. So f is increasing on (−∞,−1) and (2,∞),and f is decreasing on (−1, 2).
(b) Since f changes from increasing to decreasing at x = −1, f(−1) = 7 is a local
maximum value. Since f changes from decreasing to increasing at x = 2,
f(2) = −20 is a local minimum value.
(d)
(c) f 00(x) = 6(2x− 1) ⇒ f 00(x) > 0 on 12 ,∞ and f 00(x) < 0 on −∞, 12 .
So f is concave upward on 12,∞ and concave downward on −∞, 1
2. There
is a change in concavity at x = 12
, and we have an inflection point at 12,− 13
2.
31. (a) f(x) = 2 + 2x2 − x4 ⇒ f 0(x) = 4x− 4x3 = 4x(1− x2) = 4x(1 + x)(1− x). f 0(x) > 0 ⇔ x < −1 or
0 < x < 1 and f 0(x) < 0 ⇔ −1 < x < 0 or x > 1. So f is increasing on (−∞,−1) and (0, 1) and f is decreasing
on (−1, 0) and (1,∞).
(b) f changes from increasing to decreasing at x = −1 and x = 1, so f(−1) = 3 and f(1) = 3 are local maximum values.
f changes from decreasing to increasing at x = 0, so f(0) = 2 is a local minimum value.
(c) f 00(x) = 4− 12x2 = 4(1− 3x2). f 00(x) = 0 ⇔ 1− 3x2 = 0 ⇔x2 = 1
3⇔ x = ±1/√3. f 00(x) > 0 on −1/√3, 1/√3 and f 00(x) < 0
on −∞,−1/√3 and 1/√3,∞ . So f is concave upward on
−1/√3, 1/√3 and f is concave downward on −∞,−1/√3 and
1/√3,∞ . f ±1/√3 = 2 + 2
3 − 19 =
239 . There are points of inflection
at ±1/√3, 239
.
(d)
33. (a) h(x) = (x+ 1)5 − 5x− 2 ⇒ h0(x) = 5(x+ 1)4 − 5. h0(x) = 0 ⇔ 5(x+ 1)4 = 5 ⇔ (x+ 1)4 = 1 ⇒(x+ 1)2 = 1 ⇒ x+ 1 = 1 or x+ 1 = −1 ⇒ x = 0 or x = −2. h0(x) > 0 ⇔ x < −2 or x > 0 and
h0(x) < 0 ⇔ −2 < x < 0. So h is increasing on (−∞,−2) and (0,∞) and h is decreasing on (−2, 0).
SECTION 4.3 HOW DERIVATIVES AFFECT THE SHAPE OF A GRAPH ¤ 123
(b) h(−2) = 7 is a local maximum value and h(0) = −1 is a local minimum value. (d)
(c) h00(x) = 20(x+ 1)3 = 0 ⇔ x = −1. h00(x) > 0 ⇔ x > −1 and
h00(x) < 0 ⇔ x < −1, so h is CU on (−1,∞) and h is CD on (−∞,−1).There is a point of inflection at (−1, h(−1)) = (−1, 3).
35. (a) A(x) = x√x+ 3 ⇒ A0(x) = x · 1
2(x+3)−1/2+
√x+ 3 ·1 = x
2√x+ 3
+√x+ 3 =
x+ 2(x+ 3)
2√x+ 3
=3x+ 6
2√x+ 3
.
The domain of A is [−3,∞). A0(x) > 0 for x > −2 and A0(x) < 0 for −3 < x < −2, so A is increasing on (−2,∞)and decreasing on (−3,−2).
(b) A(−2) = −2 is a local minimum value. (d)
(c) A00(x) =2√x+ 3 · 3− (3x+ 6) · 1√
x+ 3
2√x+ 3
2
=6(x+ 3)− (3x+ 6)
4(x+ 3)3/2=
3x+ 12
4(x+ 3)3/2=
3(x+ 4)
4(x+ 3)3/2
A00(x) > 0 for all x > −3, so A is concave upward on (−3,∞). There is no inflection point.
37. (a) C(x) = x1/3(x+ 4) = x4/3 + 4x1/3 ⇒ C 0(x) = 43x
1/3 + 43x−2/3 = 4
3x−2/3(x+ 1) =
4(x+ 1)
33√x2
. C 0(x) > 0 if
−1 < x < 0 or x > 0 and C 0(x) < 0 for x < −1, so C is increasing on (−1,∞) and C is decreasing on (−∞,−1).
(b) C(−1) = −3 is a local minimum value. (d)
(c) C00(x) = 49x−2/3 − 8
9x−5/3 = 4
9x−5/3(x− 2) = 4(x− 2)
93√x5
.
C00(x) < 0 for 0 < x < 2 and C00(x) > 0 for x < 0 and x > 2, so C is
concave downward on (0, 2) and concave upward on (−∞, 0) and (2,∞).There are inflection points at (0, 0) and 2, 6 3
√2 ≈ (2, 7.56).
39. (a) f(θ) = 2 cos θ + cos2 θ, 0 ≤ θ ≤ 2π ⇒ f 0(θ) = −2 sin θ + 2cos θ (− sin θ) = −2 sin θ (1 + cos θ).f 0(θ) = 0 ⇔ θ = 0, π, and 2π. f 0(θ) > 0 ⇔ π < θ < 2π and f 0(θ) < 0 ⇔ 0 < θ < π. So f is increasing on
(π, 2π) and f is decreasing on (0, π).
(b) f(π) = −1 is a local minimum value.
(c) f 0(θ) = −2 sin θ (1 + cos θ) ⇒f 00(θ) = −2 sin θ (− sin θ) + (1 + cos θ)(−2 cos θ) = 2 sin2 θ − 2 cos θ − 2 cos2 θ
= 2(1− cos2 θ)− 2 cos θ − 2 cos2 θ = −4 cos2 θ − 2 cos θ + 2= −2(2 cos2 θ + cos θ − 1) = −2(2 cos θ − 1)(cos θ + 1)
Since −2(cos θ + 1) < 0 [for θ 6= π], f 00(θ) > 0 ⇒ 2 cos θ − 1 < 0 ⇒ cos θ < 12⇒ π
3< θ < 5π
3and
124 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
f 00(θ) < 0 ⇒ cos θ > 12 ⇒ 0 < θ < π
3 or 5π3 < θ < 2π. So f is CU on π
3 ,5π3
and f is CD on 0, π3
and
5π3, 2π . There are points of inflection at π
3, f π
3= π
3, 54
and 5π3, f 5π
3= 5π
3, 54
.
(d)
41. The nonnegative factors (x+ 1)2 and (x− 6)4 do not affect the sign of f 0(x) = (x+ 1)2(x− 3)5(x− 6)4.
So f 0(x) > 0 ⇒ (x− 3)5 > 0 ⇒ x− 3 > 0 ⇒ x > 3. Thus, f is increasing on the interval (3,∞).43. (a) From the graph, we get an estimate of f(1) ≈ 1.41 as a local maximum
value, and no local minimum value.
f(x) =x+ 1√x2 + 1
⇒ f 0(x) =1− x
(x2 + 1)3/2.
f 0(x) = 0 ⇔ x = 1. f(1) = 2√2=√2 is the exact value.
(b) From the graph in part (a), f increases most rapidly somewhere between x = − 12
and x = − 14
. To find the exact value,
we need to find the maximum value of f 0, which we can do by finding the critical numbers of f 0.
f 00(x) =2x2 − 3x− 1(x2 + 1)5/2
= 0 ⇔ x =3±√17
4. x =
3+√17
4corresponds to the minimum value of f 0.
The maximum value of f 0 occurs at x = 3−√174
≈ −0.28.
45. f(x) = cosx+ 12cos 2x ⇒ f 0(x) = − sinx− sin 2x ⇒ f 00(x) = − cosx− 2 cos 2x
(a) From the graph of f , it seems that f is CD on (0, 1), CU on (1, 2.5), CD on
(2.5, 3.7), CU on (3.7, 5.3), and CD on (5.3, 2π). The points of inflection
appear to be at (1, 0.4), (2.5,−0.6), (3.7,−0.6), and (5.3, 0.4).
(b) From the graph of f 00 (and zooming in near the zeros), it seems that f is CD
on (0, 0.94), CU on (0.94, 2.57), CD on (2.57, 3.71), CU on (3.71, 5.35),
and CD on (5.35, 2π). Refined estimates of the inflection points are
(0.94, 0.44), (2.57,−0.63), (3.71,−0.63), and (5.35, 0.44).
47. In Maple, we define f and then use the command
plot(diff(diff(f,x),x),x=-2..2);. In Mathematica, we define f
and then use Plot[Dt[Dt[f,x],x],{x,-2,2}]. We see that f 00 > 0 for
x < −0.6 and x > 0.0 [≈ 0.03] and f 00 < 0 for −0.6 < x < 0.0. So f is CU
on (−∞,−0.6) and (0.0,∞) and CD on (−0.6, 0.0).
SECTION 4.3 HOW DERIVATIVES AFFECT THE SHAPE OF A GRAPH ¤ 125
49. (a) The rate of increase of the population is initially very small, then gets larger until it reaches a maximum at about
t = 8 hours, and decreases toward 0 as the population begins to level off.
(b) The rate of increase has its maximum value at t = 8 hours.
(c) The population function is concave upward on (0, 8) and concave downward on (8, 18).
(d) At t = 8, the population is about 350, so the inflection point is about (8, 350).
51. Most students learn more in the third hour of studying than in the eighth hour, so K(3)−K(2) is larger than K(8)−K(7).
In other words, as you begin studying for a test, the rate of knowledge gain is large and then starts to taper off, so K0(t)
decreases and the graph of K is concave downward.
53. f(x) = ax3 + bx2 + cx+ d ⇒ f 0(x) = 3ax2 + 2bx+ c.
We are given that f(1) = 0 and f(−2) = 3, so f(1) = a+ b+ c+ d = 0 and
f(−2) = −8a+ 4b− 2c+ d = 3. Also f 0(1) = 3a+ 2b+ c = 0 and
f 0(−2) = 12a− 4b+ c = 0 by Fermat’s Theorem. Solving these four equations, we get
a = 29
, b = 13
, c = − 43
, d = 79
, so the function is f(x) = 192x3 + 3x2 − 12x+ 7 .
55. Suppose that f is differentiable on an interval I and f 0(x) > 0 for all x in I except x = c. To show that f is increasing on I,
let x1, x2 be two numbers in I with x1 < x2.
Case 1 x1 < x2 < c. Let J be the interval {x ∈ I | x < c}. By applying the Increasing/Decreasing Test to f
on J , we see that f is increasing on J , so f(x1) < f(x2).
Case 2 c < x1 < x2. Apply the Increasing/Decreasing Test to f on K = {x ∈ I | x > c}.
Case 3 x1 < x2 = c. Apply the proof of the Increasing/Decreasing Test, using the Mean Value Theorem (MVT)
on the interval [x1, x2] and noting that the MVT does not require f to be differentiable at the endpoints
of [x1, x2].
Case 4 c = x1 < x2. Same proof as in Case 3.
Case 5 x1 < c < x2. By Cases 3 and 4, f is increasing on [x1, c] and on [c, x2], so f(x1) < f(c) < f(x2).
In all cases, we have shown that f(x1) < f(x2). Since x1, x2 were any numbers in I with x1 < x2, we have shown that f is
increasing on I.
57. (a) Since f and g are positive, increasing, and CU on I with f 00 and g00 never equal to 0, we have f > 0, f 0 ≥ 0, f 00 > 0,
g > 0, g0 ≥ 0, g00 > 0 on I. Then (fg)0 = f 0g + fg0 ⇒ (fg)00 = f 00g + 2f 0g0 + fg00 ≥ f 00g + fg00 > 0 on I ⇒fg is CU on I.
(b) In part (a), if f and g are both decreasing instead of increasing, then f 0 ≤ 0 and g0 ≤ 0 on I, so we still have 2f 0g0 ≥ 0on I. Thus, (fg)00 = f 00g + 2f 0g0 + fg00 ≥ f 00g + fg00 > 0 on I ⇒ fg is CU on I as in part (a).
126 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
(c) Suppose f is increasing and g is decreasing [with f and g positive and CU]. Then f 0 ≥ 0 and g0 ≤ 0 on I, so 2f 0g0 ≤ 0on I and the argument in parts (a) and (b) fails.
Example 1. I = (0,∞), f(x) = x3, g(x) = 1/x. Then (fg)(x) = x2, so (fg)0(x) = 2x and
(fg)00(x) = 2 > 0 on I. Thus, fg is CU on I.
Example 2. I = (0,∞), f(x) = 4x√x, g(x) = 1/x. Then (fg)(x) = 4√x, so (fg)0(x) = 2/
√x and
(fg)00(x) = −1/√x3 < 0 on I. Thus, fg is CD on I.
Example 3. I = (0,∞), f(x) = x2, g(x) = 1/x. Thus, (fg)(x) = x, so fg is linear on I.
59. f(x) = tanx− x ⇒ f 0(x) = sec2 x− 1 > 0 for 0 < x < π2
since sec2 x > 1 for 0 < x < π2
. So f is increasing
on 0, π2
. Thus, f(x) > f(0) = 0 for 0 < x < π2 ⇒ tanx− x > 0 ⇒ tanx > x for 0 < x < π
2 .
61. Let the cubic function be f(x) = ax3 + bx2 + cx+ d ⇒ f 0(x) = 3ax2 + 2bx+ c ⇒ f 00(x) = 6ax+ 2b.
So f is CU when 6ax+ 2b > 0 ⇔ x > −b/(3a), CD when x < −b/(3a), and so the only point of inflection occurs when
x = −b/(3a). If the graph has three x-intercepts x1, x2 and x3, then the expression for f(x) must factor as
f(x) = a(x− x1)(x− x2)(x− x3). Multiplying these factors together gives us
f(x) = a[x3 − (x1 + x2 + x3)x2 + (x1x2 + x1x3 + x2x3)x− x1x2x3]
Equating the coefficients of the x2-terms for the two forms of f gives us b = −a(x1 + x2 + x3). Hence, the x-coordinate of
the point of inflection is − b
3a= −−a(x1 + x2 + x3)
3a=
x1 + x2 + x33
.
63. By hypothesis g = f 0 is differentiable on an open interval containing c. Since (c, f(c)) is a point of inflection, the concavity
changes at x = c, so f 00(x) changes signs at x = c. Hence, by the First Derivative Test, f 0 has a local extremum at x = c.
Thus, by Fermat’s Theorem f 00(c) = 0.
65. Using the fact that |x| = √x2, we have that g(x) = x |x| = x√x2 ⇒ g0(x) =
√x2 +
√x2 = 2
√x2 = 2 |x| ⇒
g00(x) = 2x x2−1/2
=2x
|x| < 0 for x < 0 and g00(x) > 0 for x > 0, so (0, 0) is an inflection point. But g00(0) does not
exist.
67. (a) f(x) = x4 sin1
x⇒ f 0(x) = x4 cos
1
x− 1
x2+ sin
1
x(4x3) = 4x3 sin
1
x− x2 cos
1
x.
g(x) = x4 2 + sin1
x= 2x4 + f(x) ⇒ g0(x) = 8x3 + f 0(x).
h(x) = x4 −2 + sin 1x
= −2x4 + f(x) ⇒ h0(x) = −8x3 + f 0(x).
It is given that f(0) = 0, so f 0(0) = limx→0
f(x)− f(0)
x− 0 = limx→0
x4 sin1
x− 0
x= lim
x→0x3 sin
1
x. Since
SECTION 4.4 LIMITS AT INFINITY; HORIZONTAL ASYMPTOTES ¤ 127
− x3 ≤ x3 sin1
x≤ x3 and lim
x→0x3 = 0, we see that f 0(0) = 0 by the Squeeze Theorem. Also,
g0(0) = 8(0)3 + f 0(0) = 0 and h0(0) = −8(0)3 + f 0(0) = 0, so 0 is a critical number of f , g, and h.
For x2n =1
2nπ[n a nonzero integer], sin 1
x2n= sin 2nπ = 0 and cos 1
x2n= cos 2nπ = 1, so f 0(x2n) = −x22n < 0.
For x2n+1 =1
(2n+ 1)π, sin 1
x2n+1= sin(2n+ 1)π = 0 and cos 1
x2n+1= cos(2n+ 1)π = −1, so
f 0(x2n+1) = x22n+1 > 0. Thus, f 0 changes sign infinitely often on both sides of 0.
Next, g0(x2n) = 8x32n + f 0(x2n) = 8x32n − x22n = x22n(8x2n − 1) < 0 for x2n < 18
, but
g0(x2n+1) = 8x32n+1 + x22n+1 = x22n+1(8x2n+1 + 1) > 0 for x2n+1 > − 18
, so g0 changes sign infinitely often on both
sides of 0.
Last, h0(x2n) = −8x32n + f 0(x2n) = −8x32n − x22n = −x22n(8x2n + 1) < 0 for x2n > − 18
and
h0(x2n+1) = −8x32n+1 + x22n+1 = x22n+1(−8x2n+1 +1) > 0 for x2n+1 < 18 , so h0 changes sign infinitely often on both
sides of 0.
(b) f(0) = 0 and since sin 1x
and hence x4 sin 1x
is both positive and negative inifinitely often on both sides of 0, and
arbitrarily close to 0, f has neither a local maximum nor a local minimum at 0.
Since 2 + sin 1x≥ 1, g(x) = x4 2 + sin
1
x> 0 for x 6= 0, so g(0) = 0 is a local minimum.
Since −2 + sin 1x≤ −1, h(x) = x4 −2 + sin 1
x< 0 for x 6= 0, so h(0) = 0 is a local maximum.
4.4 Limits at Infinity; Horizontal Asymptotes
1. (a) As x becomes large, the values of f(x) approach 5.
(b) As x becomes large negative, the values of f(x) approach 3.
3. (a) limx→2
f(x) =∞ (b) limx→−1−
f(x) =∞ (c) limx→−1+
f(x) = −∞
(d) limx→∞
f(x) = 1 (e) limx→−∞
f(x) = 2 (f ) Vertical: x = −1, x = 2; Horizontal: y = 1, y = 2
5. If f(x) = x2/2x, then a calculator gives f(0) = 0, f(1) = 0.5, f(2) = 1, f(3) = 1.125, f(4) = 1, f(5) = 0.78125,
f(6) = 0.5625, f(7) = 0.3828125, f(8) = 0.25, f(9) = 0.158203125, f(10) = 0.09765625, f(20) ≈ 0.00038147,
f(50) ≈ 2.2204× 10−12, f(100) ≈ 7.8886× 10−27.
It appears that limx→∞
x2/2x = 0.
128 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
7. limx→∞
3x2 − x+ 4
2x2 + 5x− 8 = limx→∞
(3x2 − x+ 4)/x2
(2x2 + 5x− 8)/x2[divide both the numerator and denominator by x2
(the highest power of x that appears in the denominator)]
=limx→∞
(3− 1/x+ 4/x2)limx→∞
(2 + 5/x− 8/x2) [Limit Law 5]
=limx→∞
3− limx→∞
(1/x) + limx→∞
(4/x2)
limx→∞
2 + limx→∞
(5/x)− limx→∞
(8/x2)[Limit Laws 1 and 2]
=3− lim
x→∞(1/x) + 4 lim
x→∞(1/x2)
2 + 5 limx→∞
(1/x)− 8 limx→∞
(1/x2)[Limit Laws 7 and 3]
=3− 0 + 4(0)2 + 5(0)− 8(0) [Theorem 4]
=3
2
9. limx→∞
1
2x+ 3= lim
x→∞1/x
(2x+ 3)/x=
limx→∞
(1/x)
limx→∞
(2 + 3/x)=
limx→∞
(1/x)
limx→∞
2 + 3 limx→∞
(1/x)=
0
2 + 3(0)=0
2= 0
11. limx→−∞
1− x− x2
2x2 − 7 = limx→−∞
(1− x− x2)/x2
(2x2 − 7)/x2 =lim
x→−∞(1/x2 − 1/x− 1)lim
x→−∞(2− 7/x2)
=lim
x→−∞(1/x2)− lim
x→−∞(1/x)− lim
x→−∞1
limx→−∞
2− 7 limx→−∞
(1/x2)=0− 0− 12− 7(0) = −
1
2
13. Divide both the numerator and denominator by x3 (the highest power of x that occurs in the denominator).
limx→∞
x3 + 5x
2x3 − x2 + 4= lim
x→∞
x3 + 5x
x3
2x3 − x2 + 4
x3
= limx→∞
1 +5
x2
2− 1
x+4
x3
=
limx→∞
1 +5
x2
limx→∞
2− 1
x+4
x3
=limx→∞
1 + 5 limx→∞
1
x2
limx→∞
2− limx→∞
1
x+ 4 lim
x→∞1
x3
=1 + 5(0)
2− 0 + 4(0) =1
2
15. First, multiply the factors in the denominator. Then divide both the numerator and denominator by u4.
limu→∞
4u4 + 5
(u2 − 2)(2u2 − 1) = limu→∞
4u4 + 5
2u4 − 5u2 + 2 = limu→∞
4u4 + 5
u4
2u4 − 5u2 + 2u4
= limu→∞
4 +5
u4
2− 5
u2+2
u4
=
limu→∞
4 +5
u4
limu→∞
2− 5
u2+2
u4
=limu→∞
4 + 5 limu→∞
1
u4
limu→∞
2− 5 limu→∞
1
u2+ 2 lim
u→∞1
u4
=4 + 5(0)
2− 5(0) + 2(0) =4
2= 2
SECTION 4.4 LIMITS AT INFINITY; HORIZONTAL ASYMPTOTES ¤ 129
17. limx→∞
√9x6 − x
x3 + 1= lim
x→∞
√9x6 − x /x3
(x3 + 1)/x3=
limx→∞
(9x6 − x)/x6
limx→∞
(1 + 1/x3)[since x3 =
√x6 for x > 0]
=limx→∞
9− 1/x5limx→∞
1 + limx→∞
(1/x3)=
limx→∞
9− limx→∞
(1/x5)
1 + 0
=√9− 0 = 3
19. limx→∞
√9x2 + x− 3x = lim
x→∞
√9x2 + x− 3x √
9x2 + x+ 3x√9x2 + x+ 3x
= limx→∞
√9x2 + x
2 − (3x)2√9x2 + x+ 3x
= limx→∞
9x2 + x − 9x2√9x2 + x+ 3x
= limx→∞
x√9x2 + x+ 3x
· 1/x1/x
= limx→∞
x/x
9x2/x2 + x/x2 + 3x/x= lim
x→∞1
9 + 1/x+ 3=
1√9 + 3
=1
3 + 3=1
6
21. limx→∞
√x2 + ax−√x2 + bx = lim
x→∞
√x2 + ax−√x2 + bx
√x2 + ax+
√x2 + bx√
x2 + ax+√x2 + bx
= limx→∞
(x2 + ax)− (x2 + bx)√x2 + ax+
√x2 + bx
= limx→∞
[(a− b)x]/x√x2 + ax+
√x2 + bx /
√x2
= limx→∞
a− b
1 + a/x+ 1 + b/x=
a− b√1 + 0 +
√1 + 0
=a− b
2
23. limx→∞
x+ x3 + x5
1− x2 + x4= lim
x→∞(x+ x3 + x5)/x4
(1− x2 + x4)/x4[divide by the highest power of x in the denominator]
= limx→∞
1/x3 + 1/x+ x
1/x4 − 1/x2 + 1 =∞
because (1/x3 + 1/x+ x)→∞ and (1/x4 − 1/x2 + 1)→ 1 as x→∞.
25. limx→−∞
(x4 + x5) = limx→−∞
x5( 1x+ 1) [factor out the largest power of x] = −∞ because x5 →−∞ and 1/x+ 1→ 1
as x→−∞.
Or: limx→−∞
x4 + x5 = limx→−∞
x4 (1 + x) = −∞.
27. limx→∞
x−√x = lim
x→∞
√x√x− 1 =∞ since
√x→∞ and
√x− 1→∞ as x→∞.
29. If t = 1
x, then lim
x→∞x sin
1
x= lim
t→0+
1
tsin t = lim
t→0+
sin t
t= 1.
31. (a)
From the graph of f(x) =√x2 + x+ 1 + x, we estimate
the value of limx→−∞
f(x) to be −0.5.
(b)x f(x)
−10,000 −0.4999625−100,000 −0.4999962−1,000,000 −0.4999996
From the table, we estimate the limit
to be −0.5.
130 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
(c) limx→−∞
√x2 + x+ 1 + x = lim
x→−∞√x2 + x+ 1 + x
√x2 + x+ 1− x√x2 + x+ 1− x
= limx→−∞
x2 + x+ 1 − x2√x2 + x+ 1− x
= limx→−∞
(x+ 1)(1/x)√x2 + x+ 1− x (1/x)
= limx→−∞
1 + (1/x)
− 1 + (1/x) + (1/x2)− 1=
1 + 0
−√1 + 0 + 0− 1 = −1
2
Note that for x < 0, we have√x2 = |x| = −x, so when we divide the radical by x, with x < 0, we get
1
x
√x2 + x+ 1 = − 1√
x2
√x2 + x+ 1 = − 1 + (1/x) + (1/x2).
33. limx→∞
2x+ 1
x− 2 = limx→∞
2x+ 1
xx− 2x
= limx→∞
2 +1
x
1− 2
x
=
limx→∞
2 +1
x
limx→∞
1− 2
x
=limx→∞
2 + limx→∞
1
x
limx→∞
1− limx→∞
2
x
=2 + 0
1− 0 = 2, so y = 2 is a horizontal asymptote.
The denominator x− 2 is zero when x = 2 and the numerator is not zero, so we
investigate y = f(x) =2x+ 1
x− 2 as x approaches 2. limx→2−
f(x) = −∞ because as
x→ 2− the numerator is positive and the denominator approaches 0 through
negative values. Similarly, limx→2+
f(x) =∞. Thus, x = 2 is a vertical asymptote.
The graph confirms our work.
35. limx→∞
2x2 + x− 1x2 + x− 2 = lim
x→∞
2x2 + x− 1x2
x2 + x− 2x2
= limx→∞
2 +1
x− 1
x2
1 +1
x− 2
x2
=
limx→∞
2 +1
x− 1
x2
limx→∞
1 +1
x− 2
x2
=limx→∞
2 + limx→∞
1
x− lim
x→∞1
x2
limx→∞
1 + limx→∞
1
x− 2 lim
x→∞1
x2
=2 + 0− 01 + 0− 2(0) = 2, so y = 2 is a horizontal asymptote.
y = f(x) =2x2 + x− 1x2 + x− 2 =
(2x− 1)(x+ 1)(x+ 2)(x− 1) , so lim
x→−2−f(x) =∞,
limx→−2+
f(x) = −∞, limx→1−
f(x) = −∞, and limx→1+
f(x) =∞. Thus, x = −2
and x = 1 are vertical asymptotes. The graph confirms our work.
37. y = f(x) =x3 − x
x2 − 6x+ 5 =x(x2 − 1)
(x− 1)(x− 5) =x(x+ 1)(x− 1)(x− 1)(x− 5) =
x(x+ 1)
x− 5 = g(x) for x 6= 1.
The graph of g is the same as the graph of f with the exception of a hole in the
graph of f at x = 1. By long division, g(x) = x2 + x
x− 5 = x+ 6 +30
x− 5 .
As x→ ±∞, g(x)→ ±∞, so there is no horizontal asymptote. The denominator
of g is zero when x = 5. limx→5−
g(x) = −∞ and limx→5+
g(x) =∞, so x = 5 is a
vertical asymptote. The graph confirms our work.
SECTION 4.4 LIMITS AT INFINITY; HORIZONTAL ASYMPTOTES ¤ 131
39. From the graph, it appears y = 1 is a horizontal asymptote.
limx→∞
3x3 + 500x2
x3 + 500x2 + 100x+ 2000= lim
x→∞
3x3 + 500x2
x3
x3 + 500x2 + 100x+ 2000
x3
= limx→∞
3 + (500/x)
1 + (500/x) + (100/x2) + (2000/x3)
=3 + 0
1 + 0 + 0 + 0= 3, so y = 3 is a horizontal asymptote.
The discrepancy can be explained by the choice of the viewing window. Try
[−100,000, 100,000] by [−1, 4] to get a graph that lends credibility to our
calculation that y = 3 is a horizontal asymptote.
41. Let’s look for a rational function.
(1) limx→±∞
f(x) = 0 ⇒ degree of numerator < degree of denominator
(2) limx→0
f(x) = −∞ ⇒ there is a factor of x2 in the denominator (not just x, since that would produce a sign
change at x = 0), and the function is negative near x = 0.
(3) limx→3−
f(x) =∞ and limx→3+
f(x) = −∞ ⇒ vertical asymptote at x = 3; there is a factor of (x− 3) in the
denominator.
(4) f(2) = 0 ⇒ 2 is an x-intercept; there is at least one factor of (x− 2) in the numerator.
Combining all of this information and putting in a negative sign to give us the desired left- and right-hand limits gives us
f(x) =2− x
x2(x− 3) as one possibility.
43. y = 1− x
1 + xhas domain (−∞,−1) ∪ (−1,∞).
limx→±∞
1− x
1 + x= lim
x→±∞1/x− 11/x+ 1
=0− 10 + 1
= −1, so y = −1 is a HA.
The line x = −1 is a VA.
y0 =(1 + x)(−1)− (1− x)(1)
(1 + x)2=
−2(1 + x)2
< 0 for x 6= 1. Thus,
(−∞,−1) and (−1,∞) are intervals of decrease.
y00 = −2 · −2(1 + x)
[(1 + x)2]2=
4
(1 + x)3< 0 for x < −1 and y00 > 0 for x > −1, so the curve is CD on (−∞,−1) and CU on
(−1,∞). Since x = −1 is not in the domain, there is no IP.
45. limx→±∞
x
x2 + 1= lim
x→±∞1/x
1 + 1/x2=
0
1 + 0= 0, so y = 0 is a
horizontal asymptote.
y0 =x2 + 1− x(2x)
(x2 + 1)2=
1− x2
(x2 + 1)2= 0 when x = ±1 and y0 > 0 ⇔
x2 < 1 ⇔ −1 < x < 1, so y is increasing on (−1, 1) and decreasing
on (−∞,−1) and (1,∞).[continued]
132 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
y00 =(1 + x2)2(−2x)− (1− x2)2(x2 + 1)2x
(1 + x2)4=2x(x2 − 3)(1 + x2)3
> 0 ⇔ x >√3 or −√3 < x < 0, so y is CU on
√3,∞ and −√3, 0 and CD on −∞,−√3 and 0,
√3 .
47. y = f(x) = x4 − x6 = x4(1− x2) = x4(1 + x)(1− x). The y-intercept is
f(0) = 0. The x-intercepts are 0, −1, and 1 [found by solving f(x) = 0 for x].
Since x4 > 0 for x 6= 0, f doesn’t change sign at x = 0. The function does change
sign at x = −1 and x = 1. As x→ ±∞, f(x) = x4(1− x2) approaches−∞
because x4 →∞ and (1− x2)→−∞.
49. y = f(x) = (3− x)(1 + x)2(1− x)4. The y-intercept is f(0) = 3(1)2(1)4 = 3.
The x-intercepts are 3, −1, and 1. There is a sign change at 3, but not at −1 and 1.
When x is large positive, 3− x is negative and the other factors are positive, so
limx→∞
f(x) = −∞. When x is large negative, 3− x is positive, so
limx→−∞
f(x) =∞.
51. First we plot the points which are known to be on the graph: (2,−1) and
(0, 0). We can also draw a short line segment of slope 0 at x = 2, since
we are given that f 0(2) = 0. Now we know that f 0(x) < 0 (that is, the
function is decreasing) on (0, 2), and that f 00(x) < 0 on (0, 1) and
f 00(x) > 0 on (1, 2). So we must join the points (0, 0) and (2,−1) in
such a way that the curve is concave down on (0, 1) and concave up on (1, 2). The curve must be concave up and increasing
on (2, 4) and concave down and increasing toward y = 1 on (4,∞). Now we just need to reflect the curve in the y-axis, since
we are given that f is an even function [the condition that f(−x) = f(x) for all x].
53. We are given that f(1) = f 0(1) = 0. So we can draw a short horizontal line at the point (1, 0) to represent this situation. We
are given that x = 0 and x = 2 are vertical asymptotes, with limx→0
f(x) = −∞, limx→2+
f(x) =∞ and limx→2−
f(x) = −∞, so
we can draw the parts of the curve which approach these asymptotes.
On the interval (−∞, 0), the graph is concave down, and f(x)→∞ as
x→−∞. Between the asymptotes the graph is concave down. On the
interval (2,∞) the graph is concave up, and f(x)→ 0 as x→∞, so
y = 0 is a horizontal asymptote. The diagram shows one possible function
satisfying all of the given conditions.
SECTION 4.4 LIMITS AT INFINITY; HORIZONTAL ASYMPTOTES ¤ 133
55. (a) Since −1 ≤ sinx ≤ 1 for all x, − 1x≤ sinx
x≤ 1
xfor x > 0. As x→∞, −1/x→ 0 and 1/x→ 0, so by the Squeeze
Theorem, (sinx)/x→ 0. Thus, limx→∞
sinx
x= 0.
(b) From part (a), the horizontal asymptote is y = 0. The function
y = (sinx)/x crosses the horizontal asymptote whenever sinx = 0;
that is, at x = πn for every integer n. Thus, the graph crosses the
asymptote an infinite number of times.
57. Divide the numerator and the denominator by the highest power of x in Q(x).
(a) If degP < degQ, then the numerator → 0 but the denominator doesn’t. So limx→∞
[P (x)/Q(x)] = 0.
(b) If degP > degQ, then the numerator → ±∞ but the denominator doesn’t, so limx→∞
[P (x)/Q(x)] = ±∞
(depending on the ratio of the leading coefficients of P and Q).
59. limx→∞
4x− 1x
= limx→∞
4− 1
x= 4 and lim
x→∞4x2 + 3x
x2= lim
x→∞4 +
3
x= 4. Therefore, by the Squeeze Theorem,
limx→∞
f(x) = 4.
61. Let g(x) = 3x2 + 1
2x2 + x+ 1and f(x) = |g(x)− 1.5|. Note that
limx→∞
g(x) = 32 and lim
x→∞f(x) = 0. We are interested in finding the
x-value at which f(x) < 0.05. From the graph, we find that x ≈ 14.804,
so we choose N = 15 (or any larger number).
63. For ε = 0.5, we need to find N such that√4x2 + 1
x+ 1− (−2) < 0.5 ⇔
−2.5 <√4x2 + 1
x+ 1< −1.5 whenever x ≤ N . We graph the three parts of this
inequality on the same screen, and see that the inequality holds for x ≤ −6.So we
choose N = −6 (or any smaller number).
For ε = 0.1, we need −2.1 <√4x2 + 1
x+ 1< −1.9 whenever x ≤ N.From the
graph, it seems that this inequality holds for x ≤ −22. So we choose N = −22(or any smaller number).
65. (a) 1/x2 < 0.0001 ⇔ x2 > 1/0.0001 = 10 000 ⇔ x > 100 (x > 0)
(b) If ε > 0 is given, then 1/x2 < ε ⇔ x2 > 1/ε ⇔ x > 1/√ε. Let N = 1/
√ε.
Then x > N ⇒ x >1√ε⇒ 1
x2− 0 =
1
x2< ε, so lim
x→∞1
x2= 0.
134 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
67. For x < 0, |1/x− 0| = −1/x. If ε > 0 is given, then −1/x < ε ⇔ x < −1/ε.
Take N = −1/ε. Then x < N ⇒ x < −1/ε ⇒ |(1/x)− 0| = −1/x < ε, so limx→−∞
(1/x) = 0.
69. Suppose that limx→∞
f(x) = L. Then for every ε > 0 there is a corresponding positive number N such that |f(x)− L| < ε
whenever x > N . If t = 1/x, then x > N ⇔ 0 < 1/x < 1/N ⇔ 0 < t < 1/N . Thus, for every ε > 0 there is a
corresponding δ > 0 (namely 1/N) such that |f(1/t)− L| < ε whenever 0 < t < δ. This proves that
limt→0+
f(1/t) = L = limx→∞
f(x).
Now suppose that limx→−∞
f(x) = L. Then for every ε > 0 there is a corresponding negative number N such that
|f(x)− L| < ε whenever x < N . If t = 1/x, then x < N ⇔ 1/N < 1/x < 0 ⇔ 1/N < t < 0. Thus, for every
ε > 0 there is a corresponding δ > 0 (namely −1/N) such that |f(1/t)− L| < ε whenever −δ < t < 0. This proves that
limt→0−
f(1/t) = L = limx→−∞
f(x).
4.5 Summary of Curve Sketching
1. y = f(x) = x3 + x = x(x2 + 1) A. f is a polynomial, so D = R.
B. x-intercept = 0, y-intercept = f(0) = 0 C. f(−x) = −f(x), so f is
odd; the curve is symmetric about the origin. D. f is a polynomial, so there is
no asymptote. E. f 0(x) = 3x2 + 1 > 0, so f is increasing on (−∞,∞).F. There is no critical number and hence, no local maximum or minimum value.
G. f 00(x) = 6x > 0 on (0,∞) and f 00(x) < 0 on (−∞, 0), so f is CU on
(0,∞) and CD on (−∞, 0). Since the concavity changes at x = 0, there is an
inflection point at (0, 0).
H.
3. y = f(x) = 2− 15x+ 9x2 − x3 = −(x− 2) x2 − 7x+ 1 A. D = R B. y-intercept: f(0) = 2; x-intercepts:
f(x) = 0 ⇒ x = 2 or (by the quadratic formula) x = 7±√452 ≈ 0.15, 6.85 C. No symmetry D. No asymptote
E. f 0(x) = −15 + 18x− 3x2 = −3(x2 − 6x+ 5)= −3(x− 1)(x− 5) > 0 ⇔ 1 < x < 5
so f is increasing on (1, 5) and decreasing on (−∞, 1) and (5,∞).F. Local maximum value f(5) = 27, local minimum value f(1) = −5G. f 00(x) = 18− 6x = −6(x− 3) > 0 ⇔ x < 3, so f is CU on (−∞, 3)
and CD on (3,∞). IP at (3, 11)
H.
SECTION 4.5 SUMMARY OF CURVE SKETCHING ¤ 135
5. y = f(x) = x4 + 4x3 = x3(x+ 4) A. D = R B. y-intercept: f(0) = 0;
x-intercepts: f(x) = 0 ⇔ x = −4, 0 C. No symmetry
D. No asymptote E. f 0(x) = 4x3 + 12x2 = 4x2(x+ 3) > 0 ⇔x > −3, so f is increasing on (−3,∞) and decreasing on (−∞,−3).F. Local minimum value f(−3) = −27, no local maximum
G. f 00(x) = 12x2 + 24x = 12x(x+ 2) < 0 ⇔ −2 < x < 0, so f is CD
on (−2, 0) and CU on (−∞,−2) and (0,∞). IP at (0, 0) and (−2,−16)
H.
7. y = f(x) = 2x5 − 5x2 + 1 A. D = R B. y-intercept: f(0) = 1 C. No symmetry D. No asymptote
E. f 0(x) = 10x4 − 10x = 10x(x3 − 1) = 10x(x− 1)(x2 + x+ 1), so f 0(x) < 0 ⇔ 0 < x < 1 and f 0(x) > 0 ⇔x < 0 or x > 1. Thus, f is increasing on (−∞, 0) and (1,∞) and decreasing on (0, 1). F. Local maximum value f(0) = 1,
local minimum value f(1) = −2 G. f 00(x) = 40x3 − 10 = 10(4x3 − 1)so f 00(x) = 0 ⇔ x = 1/ 3
√4. f 00(x) > 0 ⇔ x > 1/ 3
√4 and
f 00(x) < 0 ⇔ x < 1/ 3√4, so f is CD on −∞, 1/ 3
√4 and CU
on 1/ 3√4,∞ . IP at 1
3√4, 1− 9
2 3√4
2 ≈ (0.630,−0.786)
H.
9. y = f(x) = x/(x− 1) A. D = {x | x 6= 1} = (−∞, 1) ∪ (1,∞) B. x-intercept = 0, y-intercept = f(0) = 0
C. No symmetry D. limx→±∞
x
x− 1 = 1, so y = 1 is a HA. limx→1−
x
x− 1 = −∞, limx→1+
x
x− 1 =∞, so x = 1 is a VA.
E. f 0(x) = (x− 1)− x
(x− 1)2 =−1
(x− 1)2 < 0 for x 6= 1, so f is
decreasing on (−∞, 1) and (1,∞) . F. No extreme values
G. f 00(x) = 2
(x− 1)3 > 0 ⇔ x > 1, so f is CU on (1,∞) and
CD on (−∞, 1). No IP
H.
11. y = f(x) = 1/(x2 − 9) A. D = {x | x 6= ±3} = (−∞,−3) ∪ (−3, 3) ∪ (3,∞) B. y-intercept = f(0) = − 19
, no
x-intercept C. f(−x) = f(x) ⇒ f is even; the curve is symmetric about the y-axis. D. limx→±∞
1
x2 − 9 = 0, so y = 0
is a HA. limx→3−
1
x2 − 9 = −∞, limx→3+
1
x2 − 9 =∞, limx→−3−
1
x2 − 9 =∞, limx→−3+
1
x2 − 9 = −∞, so x = 3 and x = −3
are VA. E. f 0(x) = − 2x
(x2 − 9)2 > 0 ⇔ x < 0 (x 6= −3) so f is increasing on (−∞,−3) and (−3, 0) and
136 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
decreasing on (0, 3) and (3,∞). F. Local maximum value f(0) = − 19 .
G. y00 = −2(x2 − 9)2 + (2x)2(x2 − 9)(2x)(x2 − 9)4 =
6(x2 + 3)
(x2 − 9)3 > 0 ⇔
x2 > 9 ⇔ x > 3 or x < −3, so f is CU on (−∞,−3) and (3,∞) and
CD on (−3, 3). No IP
H.
13. y = f(x) = x/(x2 + 9) A. D = R B. y-intercept: f(0) = 0; x-intercept: f(x) = 0 ⇔ x = 0
C. f(−x) = −f(x), so f is odd and the curve is symmetric about the origin. D. limx→±∞
[x/(x2 + 9)] = 0, so y = 0 is a
HA; no VA E. f 0(x) = (x2 + 9)(1)− x(2x)
(x2 + 9)2=
9− x2
(x2 + 9)2=(3 + x)(3− x)
(x2 + 9)2> 0 ⇔ −3 < x < 3, so f is increasing
on (−3, 3) and decreasing on (−∞,−3) and (3,∞). F. Local minimum value f(−3) = − 16 , local maximum
value f(3) = 16
f 00(x) =(x2 + 9)2(−2x)− (9− x2) · 2(x2 + 9)(2x)
[(x2 + 9)2]2=(2x)(x2 + 9)[−(x2 + 9)− 2(9− x2)]
(x2 + 9)4=2x(x2 − 27)(x2 + 9)3
= 0 ⇔ x = 0,±√27 = ±3√3
G. f 00(x) = (x2 + 9)2(−2x)− (9− x2) · 2(x2 + 9)(2x)[(x2 + 9)2]2
=(2x)(x2 + 9) −(x2 + 9)− 2(9− x2)
(x2 + 9)4
=2x(x2 − 27)(x2 + 9)3
= 0 ⇔ x = 0, ±√27 = ±3√3 H.
f 00(x) > 0 ⇔ −3√3 < x < 0 or x > 3√3, so f is CU on −3√3, 0
and 3√3,∞ , and CD on −∞,−3√3 and 0, 3
√3 . There are three
inflection points: (0, 0) and ±3√3,± 112
√3 .
15. y = f(x) =x− 1x2
A. D = {x | x 6= 0} = (−∞, 0) ∪ (0,∞) B. No y-intercept; x-intercept: f(x) = 0 ⇔ x = 1
C. No symmetry D. limx→±∞
x− 1x2
= 0, so y = 0 is a HA. limx→0
x− 1x2
= −∞, so x = 0 is a VA.
E. f 0(x) = x2 · 1− (x− 1) · 2x(x2)2
=−x2 + 2x
x4=−(x− 2)
x3, so f 0(x) > 0 ⇔ 0 < x < 2 and f 0(x) < 0 ⇔
x < 0 or x > 2. Thus, f is increasing on (0, 2) and decreasing on (−∞, 0)
and (2,∞). F. No local minimum, local maximum value f(2) = 14
.
G. f 00(x) = x3 · (−1)− [−(x− 2)] · 3x2(x3)2
=2x3 − 6x2
x6=2(x− 3)
x4.
f 00(x) is negative on (−∞, 0) and (0, 3) and positive on (3,∞), so f is CD
on (−∞, 0) and (0, 3) and CU on (3,∞). IP at 3, 29
H.
SECTION 4.5 SUMMARY OF CURVE SKETCHING ¤ 137
17. y = f(x) =x2
x2 + 3=(x2 + 3)− 3
x2 + 3= 1− 3
x2 + 3A. D = R B. y-intercept: f(0) = 0;
x-intercepts: f(x) = 0 ⇔ x = 0 C. f(−x) = f(x), so f is even; the graph is symmetric about the y-axis.
D. limx→±∞
x2
x2 + 3= 1, so y = 1 is a HA. No VA. E. Using the Reciprocal Rule, f 0(x) = −3 · −2x
(x2 + 3)2=
6x
(x2 + 3)2.
f 0(x) > 0 ⇔ x > 0 and f 0(x) < 0 ⇔ x < 0, so f is decreasing on (−∞, 0) and increasing on (0,∞).F. Local minimum value f(0) = 0, no local maximum.
G. f 00(x) = (x2 + 3)2 · 6− 6x · 2(x2 + 3) · 2x[(x2 + 3)2]2
=6(x2 + 3)[(x2 + 3)− 4x2]
(x2 + 3)4=6(3− 3x2)(x2 + 3)3
=−18(x+ 1)(x− 1)
(x2 + 3)3
f 00(x) is negative on (−∞,−1) and (1,∞) and positive on (−1, 1),so f is CD on (−∞,−1) and (1,∞) and CU on (−1, 1). IP at ±1, 1
4
H.
19. y = f(x) = x√5− x A. The domain is {x | 5− x ≥ 0} = (−∞, 5] B. y-intercept: f(0) = 0;
x-intercepts: f(x) = 0 ⇔ x = 0, 5 C. No symmetry D. No asymptote
E. f 0(x) = x · 12 (5− x)−1/2 (−1) + (5− x)1/2 · 1 = 12 (5− x)−1/2 [−x+ 2(5− x)] =
10− 3x2√5− x
> 0 ⇔
x < 103
, so f is increasing on −∞, 103
and decreasing on 103, 5 .
F. Local maximum value f 103= 10
9
√15 ≈ 4.3; no local minimum
G. f 00(x) =2(5− x)1/2(−3)− (10− 3x) · 2 1
2(5− x)−1/2(−1)
2√5− x
2
=(5− x)−1/2[−6(5− x) + (10− 3x)]
4(5− x)=
3x− 204(5− x)3/2
f 00(x) < 0 for x < 5, so f is CD on (−∞, 5). No IP
H.
21. y = f(x) =√x2 + x− 2 = (x+ 2)(x− 1) A. D = {x | (x+ 2)(x− 1) ≥ 0} = (−∞,−2] ∪ [1,∞)
B. y-intercept: none; x-intercepts: −2 and 1 C. No symmetry D. No asymptote
E. f 0(x) = 12(x2 + x− 2)−1/2(2x+ 1) = 2x+ 1
2√x2 + x− 2 , f 0(x) = 0 if x = − 1
2, but − 1
2is not in the domain.
f 0(x) > 0 ⇒ x > − 12
and f 0(x) < 0 ⇒ x < − 12
, so (considering the domain) f is increasing on (1,∞) and
f is decreasing on (−∞,−2). F. No local extrema
G. f 00(x) =2(x2 + x− 2)1/2(2)− (2x+ 1) · 2 · 1
2(x2 + x− 2)−1/2(2x+ 1)
2√x2 + x− 2 2
=(x2 + x− 2)−1/2 4(x2 + x− 2)− (4x2 + 4x+ 1)
4(x2 + x− 2)
=−9
4(x2 + x− 2)3/2 < 0
so f is CD on (−∞,−2) and (1,∞). No IP
H.
138 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
23. y = f(x) = x/√x2 + 1 A. D = R B. y-intercept: f(0) = 0; x-intercepts: f(x) = 0 ⇒ x = 0
C. f(−x) = −f(x), so f is odd; the graph is symmetric about the origin.
D. limx→∞
f(x) = limx→∞
x√x2 + 1
= limx→∞
x/x√x2 + 1/x
= limx→∞
x/x√x2 + 1/
√x2= lim
x→∞1
1 + 1/x2=
1√1 + 0
= 1
and
limx→−∞
f(x) = limx→−∞
x√x2 + 1
= limx→−∞
x/x√x2 + 1/x
= limx→−∞
x/x√x2 + 1/ −√x2
= limx→−∞
1
− 1 + 1/x2
=1
−√1 + 0 = −1 so y = ±1 are HA.
No VA.
E. f 0(x) =
√x2 + 1− x · 2x
2√x2 + 1
[(x2 + 1)1/2]2=
x2 + 1− x2
(x2 + 1)3/2=
1
(x2 + 1)3/2> 0 for all x, so f is increasing on R.
F. No extreme values
G. f 00(x) = − 32(x2 + 1)−5/2 · 2x = −3x
(x2 + 1)5/2, so f 00(x) > 0 for x < 0
and f 00(x) < 0 for x > 0. Thus, f is CU on (−∞, 0) and CD on (0,∞).IP at (0, 0)
H.
25. y = f(x) =√1− x2/x A. D = {x | |x| ≤ 1, x 6= 0} = [−1, 0) ∪ (0, 1] B. x-intercepts ±1, no y-intercept
C. f(−x) = −f(x), so the curve is symmetric about (0, 0) . D. limx→0+
√1− x2
x=∞, lim
x→0−
√1− x2
x= −∞,
so x = 0 is a VA. E. f 0(x) =−x2/√1− x2 −√1− x2
x2= − 1
x2√1− x2
< 0, so f is decreasing
on (−1, 0) and (0, 1). F. No extreme values
G. f 00(x) =2− 3x2
x3(1− x2)3/2> 0 ⇔ −1 < x < − 2
3or 0 < x < 2
3, so
f is CU on −1,− 23
and 0, 23
and CD on − 23, 0 and 2
3, 1 .
IP at ± 23,± 1√
2
H.
27. y = f(x) = x− 3x1/3 A. D = R B. y-intercept: f(0) = 0; x-intercepts: f(x) = 0 ⇒ x = 3x1/3 ⇒x3 = 27x ⇒ x3 − 27x = 0 ⇒ x(x2 − 27) = 0 ⇒ x = 0, ±3√3 C. f(−x) = −f(x), so f is odd;
the graph is symmetric about the origin. D. No asymptote E. f 0(x) = 1− x−2/3 = 1− 1
x2/3=
x2/3 − 1x2/3
.
f 0(x) > 0 when |x| > 1 and f 0(x) < 0 when 0 < |x| < 1, so f is increasing on (−∞,−1) and (1,∞), and
decreasing on (−1, 0) and (0, 1) [hence decreasing on (−1, 1) since f is
continuous on (−1, 1)]. F. Local maximum value f(−1) = 2, local minimum
value f(1) = −2 G. f 00(x) = 23x−5/3 < 0 when x < 0 and f 00(x) > 0
when x > 0, so f is CD on (−∞, 0) and CU on (0,∞). IP at (0, 0)
H.
SECTION 4.5 SUMMARY OF CURVE SKETCHING ¤ 139
29. y = f(x) = 3√x2 − 1 A. D = R B. y-intercept: f(0) = −1; x-intercepts: f(x) = 0 ⇔ x2 − 1 = 0 ⇔
x = ±1 C. f(−x) = f(x), so the curve is symmetric about the y-axis. D. No asymptote
E. f 0(x) = 13(x2 − 1)−2/3(2x) = 2x
3 3 (x2 − 1)2 . f 0(x) > 0 ⇔ x > 0 and f 0(x) < 0 ⇔ x < 0, so f is
increasing on (0,∞) and decreasing on (−∞, 0). F. Local minimum value f(0) = −1
G. f 00(x) = 2
3· (x
2 − 1)2/3(1)− x · 23(x2 − 1)−1/3(2x)
[(x2 − 1)2/3]2
=2
9· (x
2 − 1)−1/3[3(x2 − 1)− 4x2](x2 − 1)4/3 = − 2(x2 + 3)
9(x2 − 1)5/3
f 00(x) > 0 ⇔ −1 < x < 1 and f 00(x) < 0 ⇔ x < −1 or x > 1, so
f is CU on (−1, 1) and f is CD on (−∞,−1) and (1,∞). IP at (±1, 0)
H.
31. y = f(x) = 3 sinx− sin3 x A. D = R B. y-intercept: f(0) = 0; x-intercepts: f(x) = 0 ⇒sinx (3− sin2 x) = 0 ⇒ sinx = 0 [since sin2 x ≤ 1 < 3] ⇒ x = nπ, n an integer.
C. f(−x) = −f(x), so f is odd; the graph (shown for −2π ≤ x ≤ 2π) is symmetric about the origin and periodic
with period 2π. D. No asymptote E. f 0(x) = 3 cosx− 3 sin2 x cosx = 3cosx (1− sin2 x) = 3 cos3 x.
f 0(x) > 0 ⇔ cosx > 0 ⇔ x ∈ 2nπ − π2, 2nπ + π
2for each integer n, and f 0(x) < 0 ⇔ cosx < 0 ⇔
x ∈ 2nπ + π2, 2nπ + 3π
2for each integer n. Thus, f is increasing on 2nπ − π
2, 2nπ + π
2for each integer n,
and f is decreasing on 2nπ + π2, 2nπ + 3π
2for each integer n.
F. f has local maximum values f(2nπ + π2) = 2 and local minimum values f(2nπ + 3π
2) = −2.
G. f 00(x) = −9 sinx cos2 x = −9 sinx (1− sin2 x) = −9 sinx (1− sinx)(1 + sinx).f 00(x) < 0 ⇔ sinx > 0 and sinx 6= ±1 ⇔ x ∈ 2nπ, 2nπ + π
2∪ 2nπ + π
2, 2nπ + π for some integer n.
f 00(x) > 0 ⇔ sinx < 0 and sinx 6= ±1 ⇔ x ∈ (2n− 1)π, (2n− 1)π + π2∪ (2n− 1)π + π
2 , 2nπ
for some integer n. Thus, f is CD on the intervals 2nπ, 2n+ 12π and
2n+ 12π, (2n+ 1)π [hence CD on the intervals (2nπ, (2n+ 1)π)]
for each integer n, and f is CU on the intervals (2n− 1)π, 2n− 12π and
2n− 12π, 2nπ [hence CU on the intervals ((2n− 1)π, 2nπ)]
for each integer n. f has inflection points at (nπ, 0) for each integer n.
H.
33. y = f(x) = x tanx, −π2< x < π
2A. D = −π
2, π2
B. Intercepts are 0 C. f(−x) = f(x), so the curve is
symmetric about the y-axis. D. limx→(π/2)−
x tanx =∞ and limx→−(π/2)+
x tanx =∞, so x = π2
and x = −π2
are VA.
E. f 0(x) = tanx+ x sec2 x > 0 ⇔ 0 < x < π2
, so f increases on 0, π2
and decreases on −π2 , 0 . F. Absolute and local minimum value f(0) = 0.
G. y00 = 2 sec2 x+ 2x tanx sec2 x > 0 for −π2< x < π
2, so f is
CU on −π2, π2
. No IP
H.
140 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
35. y = f(x) = 12x− sinx, 0 < x < 3π A. D = (0, 3π) B. No y-intercept. The x-intercept, approximately 1.9, can be
found using Newton’s Method. C. No symmetry D. No asymptote E. f 0(x) = 12− cos x > 0 ⇔ cosx < 1
2⇔
π3 < x < 5π
3 or 7π3 < x < 3π, so f is increasing on π
3 ,5π3
and 7π3 , 3π and decreasing on 0, π3 and 5π
3 ,7π3
.
F. Local minimum value f π3= π
6−√32
, local maximum value
f 5π3
= 5π6 +
√32 , local minimum value f 7π
3= 7π
6 −√32
G. f 00(x) = sinx > 0 ⇔ 0 < x < π or 2π < x < 3π, so f is CU on
(0, π) and (2π, 3π) and CD on (π, 2π). IPs at π, π2
and (2π, π)
H.
37. y = f(x) =sinx
1 + cosx
⎡⎢⎣ whencosx 6= 1
=sinx
1 + cosx· 1− cosx1− cosx =
sinx (1− cosx)sin2 x
=1− cosxsinx
= cscx− cotx
⎤⎥⎦A. The domain of f is the set of all real numbers except odd integer multiples of π. B. y-intercept: f(0) = 0; x-intercepts:
x = nπ, n an even integer. C. f(−x) = −f(x), so f is an odd function; the graph is symmetric about the origin and has
period 2π. D. When n is an odd integer, limx→(nπ)−
f(x) =∞ and limx→(nπ)+
f(x) = −∞, so x = nπ is a VA for each odd
integer n. No HA. E. f 0(x) = (1 + cosx) · cosx− sinx(− sinx)(1 + cosx)2
=1+ cosx
(1 + cosx)2=
1
1 + cosx. f 0(x) > 0 for all x
except odd multiples of π, so f is increasing on ((2k − 1)π, (2k + 1)π) for each integer k. F. No extreme values
G. f 00(x) = sinx
(1 + cosx)2> 0 ⇒ sinx > 0 ⇒
x ∈ (2kπ, (2k + 1)π) and f 00(x) < 0 on ((2k − 1)π, 2kπ) for each
integer k. f is CU on (2kπ, (2k + 1)π) and CD on ((2k − 1)π, 2kπ)for each integer k. f has IPs at (2kπ, 0) for each integer k.
H.
39. m = f(v) =m0
1− v2/c2. The m-intercept is f(0) = m0. There are no v-intercepts. lim
v→c−f(v) =∞, so v = c is a VA.
f 0(v) = − 12m0(1− v2/c2)−3/2(−2v/c2) = m0v
c2(1− v2/c2)3/2=
m0v
c2(c2 − v2)3/2
c3
=m0cv
(c2 − v2)3/2> 0, so f is
increasing on (0, c). There are no local extreme values.
f 00(v) =(c2 − v2)3/2(m0c)−m0cv · 32 (c2 − v2)1/2(−2v)
[(c2 − v2)3/2]2
=m0c(c
2 − v2)1/2[(c2 − v2) + 3v2]
(c2 − v2)3=
m0c(c2 + 2v2)
(c2 − v2)5/2> 0,
so f is CU on (0, c). There are no inflection points.
SECTION 4.5 SUMMARY OF CURVE SKETCHING ¤ 141
41. y = − W
24EIx4 +
WL
12EIx3 − WL2
24EIx2 = − W
24EIx2 x2 − 2Lx+ L2
=−W24EI
x2(x− L)2 = cx2(x− L)2
where c = − W
24EIis a negative constant and 0 ≤ x ≤ L. We sketch
f(x) = cx2(x− L)2 for c = −1. f(0) = f(L) = 0.
f 0(x) = cx2[2(x− L)] + (x− L)2(2cx) = 2cx(x− L)[x+ (x− L)] = 2cx(x− L)(2x− L). So for 0 < x < L,
f 0(x) > 0 ⇔ x(x− L)(2x− L) < 0 [since c < 0] ⇔ L/2 < x < L and f 0(x) < 0 ⇔ 0 < x < L/2.
Thus, f is increasing on (L/2, L) and decreasing on (0, L/2), and there is a local and absolute
minimum at the point (L/2, f(L/2)) = L/2, cL4/16 . f 0(x) = 2c[x(x− L)(2x− L)] ⇒f 00(x) = 2c[1(x− L)(2x− L) + x(1)(2x− L) + x(x− L)(2)] = 2c(6x2 − 6Lx+ L2) = 0 ⇔
x =6L±√12L2
12= 1
2L±√36 L, and these are the x-coordinates of the two inflection points.
43. y = x2 + 1
x+ 1. Long division gives us: x− 1
x+ 1 x2 + 1
x2 + x
− x+ 1
− x− 12
Thus, y = f(x) =x2 + 1
x+ 1= x− 1 + 2
x+ 1and f(x)− (x− 1) = 2
x+ 1=
2
x
1 +1
x
[for x 6= 0] → 0 as x→ ±∞.
So the line y = x− 1 is a slant asymptote (SA).
45. y = 4x3 − 2x2 + 52x2 + x− 3 . Long division gives us: 2x− 2
2x2 + x− 3 4x3 − 2x2 + 5
4x3 + 2x2 − 6x− 4x2 + 6x+ 5− 4x2 − 2x+ 6
8x− 1
Thus, y = f(x) =4x3 − 2x2 + 52x2 + x− 3 = 2x− 2 + 8x− 1
2x2 + x− 3 and f(x)− (2x− 2) = 8x− 12x2 + x− 3 =
8
x− 1
x2
2 +1
x− 3
x2
[for x 6= 0] → 0 as x→ ±∞. So the line y = 2x− 2 is a SA.
47. y = f(x) =−2x2 + 5x− 1
2x− 1 = −x+ 2 + 1
2x− 1 A. D = x ∈ R | x 6= 12= −∞, 1
2∪ 1
2 ,∞
B. y-intercept: f(0) = 1; x-intercepts: f(x) = 0 ⇒ −2x2 + 5x− 1 = 0 ⇒ x =−5±√17−4 ⇒ x ≈ 0.22, 2.28.
C. No symmetry D. limx→(1/2)−
f(x) = −∞ and limx→(1/2)+
f(x) =∞, so x = 12
is a VA.
[continued]
142 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
limx→±∞
[f(x)− (−x+ 2)] = limx→±∞
1
2x− 1 = 0, so the line y = −x+ 2 is a SA.
E. f 0(x) = −1− 2
(2x− 1)2 < 0 for x 6= 12
, so f is decreasing on −∞, 12
and 12,∞ . F. No extreme values G. f 0(x) = −1− 2(2x− 1)−2 ⇒
f 00(x) = −2(−2)(2x− 1)−3(2) = 8
(2x− 1)3 , so f 00(x) > 0 when x > 12
and
f 00(x) < 0 when x < 12 . Thus, f is CU on 1
2 ,∞ and CD on −∞, 12 . No IP
H.
49. y = f(x) = (x2 + 4)/x = x+ 4/x A. D = {x | x 6= 0} = (−∞, 0) ∪ (0,∞) B. No intercept
C. f(−x) = −f(x) ⇒ symmetry about the origin D. limx→∞
(x+ 4/x) =∞ but f(x)− x = 4/x→ 0 as x→ ±∞,
so y = x is a slant asymptote. limx→0+
(x+ 4/x) =∞ and
limx→0−
(x+ 4/x) = −∞, so x = 0 is a VA. E. f 0(x) = 1− 4/x2 > 0 ⇔
x2 > 4 ⇔ x > 2 or x < −2, so f is increasing on (−∞,−2) and (2,∞) and
decreasing on (−2, 0) and (0, 2). F. Local maximum value f(−2) = −4, local
minimum value f(2) = 4 G. f 00(x) = 8/x3 > 0 ⇔ x > 0 so f is CU on
(0,∞) and CD on (−∞, 0). No IP
H.
51. y = f(x) =2x3 + x2 + 1
x2 + 1= 2x+ 1 +
−2xx2 + 1
A. D = R B. y-intercept: f(0) = 1; x-intercept: f(x) = 0 ⇒
0 = 2x3 + x2 + 1 = (x+ 1)(2x2 − x+ 1) ⇒ x = −1 C. No symmetry D. No VA
limx→±∞
[f(x)− (2x+ 1)] = limx→±∞
−2xx2 + 1
= limx→±∞
−2/x1 + 1/x2
= 0, so the line y = 2x+ 1 is a slant asymptote.
E. f 0(x) = 2 + (x2 + 1)(−2)− (−2x)(2x)(x2 + 1)2
=2(x4 + 2x2 + 1)− 2x2 − 2 + 4x2
(x2 + 1)2=2x4 + 6x2
(x2 + 1)2=2x2(x2 + 3)
(x2 + 1)2
so f 0(x) > 0 if x 6= 0. Thus, f is increasing on (−∞, 0) and (0,∞). Since f is continuous at 0, f is increasing on R.
F. No extreme values
G. f 00(x) =(x2 + 1)2 · (8x3 + 12x)− (2x4 + 6x2) · 2(x2 + 1)(2x)
[(x2 + 1)2]2
=4x(x2 + 1)[(x2 + 1)(2x2 + 3)− 2x4 − 6x2]
(x2 + 1)4=4x(−x2 + 3)(x2 + 1)3
so f 00(x) > 0 for x < −√3 and 0 < x <√3, and f 00(x) < 0 for
−√3 < x < 0 and x >√3. f is CU on −∞,−√3 and 0,
√3 ,
and CD on −√3, 0 and√3,∞ . There are three IPs: (0, 1),
−√3,− 32
√3 + 1 ≈ (−1.73,−1.60), and
√3, 3
2
√3 + 1 ≈ (1.73, 3.60).
H.
SECTION 4.5 SUMMARY OF CURVE SKETCHING ¤ 143
53. y = f(x) =√4x2 + 9 ⇒ f 0(x) =
4x√4x2 + 9
⇒
f 00(x) =
√4x2 + 9 · 4− 4x · 4x/√4x2 + 9
4x2 + 9=4(4x2 + 9)− 16x2(4x2 + 9)3/2
=36
(4x2 + 9)3/2. f is defined on (−∞,∞).
f(−x) = f(x), so f is even, which means its graph is symmetric about the y-axis. The y-intercept is f(0) = 3. There are no
x-intercepts since f(x) > 0 for all x.
limx→∞
√4x2 + 9− 2x = lim
x→∞
√4x2 + 9− 2x √
4x2 + 9 + 2x√4x2 + 9 + 2x
= limx→∞
(4x2 + 9)− 4x2√4x2 + 9 + 2x
= limx→∞
9√4x2 + 9 + 2x
= 0
and, similarly, limx→−∞
√4x2 + 9 + 2x = lim
x→−∞9√
4x2 + 9− 2x = 0,
so y = ±2x are slant asymptotes. f is decreasing on (−∞, 0) and increasing on (0,∞) with local minimum f(0) = 3.
f 00(x) > 0 for all x, so f is CU on R.
55. x2
a2− y2
b2= 1 ⇒ y = ± b
a
√x2 − a2. Now
limx→∞
b
a
√x2 − a2 − b
ax =
b
a· limx→∞
√x2 − a2 − x
√x2 − a2 + x√x2 − a2 + x
=b
a· limx→∞
−a2√x2 − a2 + x
= 0,
which shows that y = b
ax is a slant asymptote. Similarly,
limx→∞
− b
a
√x2 − a2 − − b
ax = − b
a· limx→∞
−a2√x2 − a2 + x
= 0, so y = − b
ax is a slant asymptote.
57. limx→±∞
f(x)− x3 = limx→±∞
x4 + 1
x− x4
x= lim
x→±∞1
x= 0, so the graph of f is asymptotic to that of y = x3.
A. D = {x | x 6= 0} B. No intercept C. f is symmetric about the origin. D. limx→0−
x3 +1
x= −∞ and
limx→0+
x3 +1
x=∞, so x = 0 is a vertical asymptote, and as shown above, the graph of f is asymptotic to that of y = x3.
E. f 0(x) = 3x2 − 1/x2 > 0 ⇔ x4 > 13⇔ |x| > 1
4√3 , so f is increasing on −∞,− 14√3
and 14√3,∞ and
decreasing on − 14√3, 0 and 0,
14√3
. F. Local maximum value
f − 14√3
= −4 · 3−5/4, local minimum value f 14√3
= 4 · 3−5/4
G. f 00(x) = 6x+ 2/x3 > 0 ⇔ x > 0, so f is CU on (0,∞) and CD
on (−∞, 0). No IP
H.
144 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
4.6 Graphing with Calculus and Calculators
1. f(x) = 4x4 − 32x3 + 89x2 − 95x+ 29 ⇒ f 0(x) = 16x3 − 96x2 + 178x− 95 ⇒ f 00(x) = 48x2 − 192x+ 178.
f(x) = 0 ⇔ x ≈ 0.5, 1.60; f 0(x) = 0 ⇔ x ≈ 0.92, 2.5, 2.58 and f 00(x) = 0 ⇔ x ≈ 1.46, 2.54.
From the graphs of f 0, we estimate that f 0 < 0 and that f is decreasing on (−∞, 0.92) and (2.5, 2.58), and that f 0 > 0 and f
is increasing on (0.92, 2.5) and (2.58,∞) with local minimum values f(0.92) ≈ −5.12 and f(2.58) ≈ 3.998 and local
maximum value f(2.5) = 4. The graphs of f 0 make it clear that f has a maximum and a minimum near x = 2.5, shown more
clearly in the fourth graph.
From the graph of f 00, we estimate that f 00 > 0 and that f is CU on
(−∞, 1.46) and (2.54,∞), and that f 00 < 0 and f is CD on (1.46, 2.54).
There are inflection points at about (1.46,−1.40) and (2.54, 3.999).
3. f(x) = x6 − 10x5 − 400x4 + 2500x3 ⇒ f 0(x) = 6x5 − 50x4 − 1600x3 + 7500x2 ⇒f 00(x) = 30x4 − 200x3 − 4800x2 + 1500x
SECTION 4.6 GRAPHING WITH CALCULUS AND CALCULATORS ¤ 145
From the graph of f 0, we estimate that f is decreasing on (−∞,−15), increasing on (−15, 4.40), decreasing
on (4.40, 18.93), and increasing on (18.93,∞), with local minimum values of f(−15) ≈ −9,700,000 and
f(18.93) ≈ −12,700,000 and local maximum value f(4.40) ≈ 53,800. From the graph of f 00, we estimate that f is CU on
(−∞,−11.34), CD on (−11.34, 0), CU on (0, 2.92), CD on (2.92, 15.08), and CU on (15.08,∞). There is an inflection
point at (0, 0) and at about (−11.34,−6,250,000), (2.92, 31,800), and (15.08,−8,150,000).
5. f(x) = x
x3 − x2 − 4x+ 1 ⇒ f 0(x) =−2x3 + x2 + 1
(x3 − x2 − 4x+ 1)2 ⇒ f 00(x) =2(3x5 − 3x4 + 5x3 − 6x2 + 3x+ 4)
(x3 − x2 − 4x+ 1)3
We estimate from the graph of f that y = 0 is a horizontal asymptote, and that there are vertical asymptotes at x = −1.7,
x = 0.24, and x = 2.46. From the graph of f 0, we estimate that f is increasing on (−∞,−1.7), (−1.7, 0.24), and (0.24, 1),
and that f is decreasing on (1, 2.46) and (2.46,∞). There is a local maximum value at f(1) = − 13
. From the graph of f 00, we
estimate that f is CU on (−∞,−1.7), (−0.506, 0.24), and (2.46,∞), and that f is CD on (−1.7,−0.506) and (0.24, 2.46).
There is an inflection point at (−0.506,−0.192).
7. f(x) = x2 − 4x+ 7cosx, −4 ≤ x ≤ 4. f 0(x) = 2x− 4− 7 sinx ⇒ f 00(x) = 2− 7 cosx.
f(x) = 0 ⇔ x ≈ 1.10; f 0(x) = 0 ⇔ x ≈ −1.49, −1.07, or 2.89; f 00(x) = 0 ⇔ x = ± cos−1 27≈ ±1.28.
From the graphs of f 0, we estimate that f is decreasing (f 0 < 0) on (−4,−1.49), increasing on (−1.49,−1.07), decreasing
on (−1.07, 2.89), and increasing on (2.89, 4), with local minimum values f(−1.49) ≈ 8.75 and f(2.89) ≈ −9.99 and local
maximum value f(−1.07) ≈ 8.79 (notice the second graph of f). From the graph of f 00, we estimate that f is CU (f 00 > 0)
on (−4,−1.28), CD on (−1.28, 1.28), and CU on (1.28, 4). There are inflection points at about (−1.28, 8.77)and (1.28,−1.48).
146 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
9. f(x) = 1 + 1
x+8
x2+1
x3⇒ f 0(x) = − 1
x2− 16
x3− 3
x4= − 1
x4(x2 + 16x+ 3) ⇒
f 00(x) =2
x3+48
x4+12
x5=2
x5(x2 + 24x+ 6).
From the graphs, it appears that f increases on (−15.8,−0.2) and decreases on (−∞,−15.8), (−0.2, 0), and (0,∞); that f
has a local minimum value of f(−15.8) ≈ 0.97 and a local maximum value of f(−0.2) ≈ 72; that f is CD on (−∞,−24)and (−0.25, 0) and is CU on (−24,−0.25) and (0,∞); and that f has IPs at (−24, 0.97) and (−0.25, 60).
To find the exact values, note that f 0 = 0 ⇒ x =−16±√256− 12
2= −8±√61 [≈ −0.19 and −15.81].
f 0 is positive (f is increasing) on −8−√61,−8 +√61 and f 0 is negative (f is decreasing) on −∞,−8−√61 ,
−8 +√61, 0 , and (0,∞). f 00 = 0 ⇒ x =−24±√576− 24
2= −12±√138 [≈ −0.25 and −23.75]. f 00 is
positive (f is CU) on −12−√138,−12 +√138 and (0,∞) and f 00 is negative (f is CD) on −∞,−12−√138
and −12 +√138, 0 .
11. f(x) =(x+ 4)(x− 3)2
x4(x− 1) has VA at x = 0 and at x = 1 since limx→0
f(x) = −∞,
limx→1−
f(x) = −∞ and limx→1+
f(x) =∞.
f(x) =
x+ 4
x· (x− 3)
2
x2
x4
x3· (x− 1)
dividing numeratorand denominator by x3
=(1 + 4/x)(1− 3/x)2
x(x− 1) → 0
as x→ ±∞, so f is asymptotic to the x-axis.
Since f is undefined at x = 0, it has no y-intercept. f(x) = 0 ⇒ (x+4)(x− 3)2 = 0 ⇒ x = −4 or x = 3, so f has
x-intercepts−4 and 3. Note, however, that the graph of f is only tangent to the x-axis and does not cross it at x = 3, since f is
positive as x→ 3− and as x→ 3+.
From these graphs, it appears that f has three maximum values and one minimum value. The maximum values are
SECTION 4.6 GRAPHING WITH CALCULUS AND CALCULATORS ¤ 147
approximately f(−5.6) = 0.0182, f(0.82) = −281.5 and f(5.2) = 0.0145 and we know (since the graph is tangent to the
x-axis at x = 3) that the minimum value is f(3) = 0.
13. f(x) = x2(x+ 1)3
(x− 2)2(x− 4)4 ⇒ f 0(x) = −x(x+ 1)2(x3 + 18x2 − 44x− 16)(x− 2)3(x− 4)5 [from CAS].
From the graphs of f 0, it seems that the critical points which indicate extrema occur at x ≈ −20, −0.3, and 2.5, as estimated
in Example 3. (There is another critical point at x = −1, but the sign of f 0 does not change there.) We differentiate again,
obtaining f 00(x) = 2(x+ 1)(x6 + 36x5 + 6x4 − 628x3 + 684x2 + 672x+ 64)
(x− 2)4(x− 4)6 .
From the graphs of f 00, it appears that f is CU on (−35.3,−5.0), (−1,−0.5), (−0.1, 2), (2, 4) and (4,∞) and CD
on (−∞,−35.3), (−5.0,−1) and (−0.5,−0.1). We check back on the graphs of f to find the y-coordinates of the
inflection points, and find that these points are approximately (−35.3,−0.015), (−5.0,−0.005), (−1, 0), (−0.5, 0.00001),
and (−0.1, 0.0000066).
15. y = f(x) =
√x
x2 + x+ 1. From a CAS, y0 = − 3x2 + x− 1
2√x (x2 + x+ 1)2
and y00 = 15x4 + 10x3 − 15x2 − 6x− 14x3/2(x2 + x+ 1)3
.
f 0(x) = 0 ⇔ x ≈ 0.43, so f is increasing on (0, 0.43) and decreasing on (0.43,∞). There is a local maximum value of
f(0.43) ≈ 0.41. f 00(x) = 0 ⇔ x ≈ 0.94, so f is CD on (0, 0.94) and CU on (0.94,∞). There is an inflection point at
(0.94, 0.34).
148 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
17. y = f(x) =√x+ 5 sinx, x ≤ 20.
From a CAS, y0 = 5cosx+ 1
2√x+ 5 sinx
and y00 = −10 cosx+ 25 sin2 x+ 10x sinx+ 26
4(x+ 5 sinx)3/2.
We’ll start with a graph of g(x) = x+ 5 sinx. Note that f(x) = g(x) is only defined if g(x) ≥ 0. g(x) = 0 ⇔ x = 0
or x ≈ −4.91, −4.10, 4.10, and 4.91. Thus, the domain of f is [−4.91,−4.10] ∪ [0, 4.10] ∪ [4.91, 20].
From the expression for y0, we see that y0 = 0 ⇔ 5 cosx+ 1 = 0 ⇒ x1 = cos−1 − 1
5≈ 1.77 and
x2 = 2π − x1 ≈ −4.51 (not in the domain of f ). The leftmost zero of f 0 is x1 − 2π ≈ −4.51. Moving to the right, the zeros
of f 0 are x1, x1 + 2π, x2 + 2π, x1 + 4π, and x2 + 4π. Thus, f is increasing on (−4.91,−4.51), decreasing on
(−4.51,−4.10), increasing on (0, 1.77), decreasing on (1.77, 4.10), increasing on (4.91, 8.06), decreasing on (8.06, 10.79),
increasing on (10.79, 14.34), decreasing on (14.34, 17.08), and increasing on (17.08, 20). The local maximum values are
f(−4.51) ≈ 0.62, f(1.77) ≈ 2.58, f(8.06) ≈ 3.60, and f(14.34) ≈ 4.39. The local minimum values are f(10.79) ≈ 2.43and f(17.08) ≈ 3.49.
f is CD on (−4.91,−4.10), (0, 4.10), (4.91, 9.60), CU on (9.60, 12.25), CD
on (12.25, 15.81), CU on (15.81, 18.65), and CD on (18.65, 20). There are
inflection points at (9.60, 2.95), (12.25, 3.27), (15.81, 3.91), and (18.65, 4.20).
19.
From the graph of f(x) = sin(x+ sin 3x) in the viewing rectangle [0, π] by [−1.2, 1.2], it looks like f has two maxima
and two minima. If we calculate and graph f 0(x) = [cos(x+ sin 3x)] (1 + 3 cos 3x) on [0, 2π], we see that
the graph of f 0 appears to be almost tangent to the x-axis at about x = 0.7. The graph of
f 00 = − [sin(x+ sin 3x)] (1 + 3 cos 3x)2 + cos(x+ sin 3x)(−9 sin 3x) is even more interesting near this x-value:
it seems to just touch the x-axis.
SECTION 4.6 GRAPHING WITH CALCULUS AND CALCULATORS ¤ 149
If we zoom in on this place on the graph of f 00, we see that f 00 actually does cross the axis twice near x = 0.65,
indicating a change in concavity for a very short interval. If we look at the graph of f 0 on the same interval, we see that it
changes sign three times near x = 0.65, indicating that what we had thought was a broad extremum at about x = 0.7 actually
consists of three extrema (two maxima and a minimum). These maximum values are roughly f(0.59) = 1 and f(0.68) = 1,
and the minimum value is roughly f(0.64) = 0.99996. There are also a maximum value of about f(1.96) = 1 and minimum
values of about f(1.46) = 0.49 and f(2.73) = −0.51. The points of inflection on (0, π) are about (0.61, 0.99998),
(0.66, 0.99998), (1.17, 0.72), (1.75, 0.77), and (2.28, 0.34). On (π, 2π), they are about (4.01,−0.34), (4.54,−0.77),(5.11,−0.72), (5.62,−0.99998), and (5.67,−0.99998). There are also IP at (0, 0) and (π, 0). Note that the function is odd
and periodic with period 2π, and it is also rotationally symmetric about all points of the form ((2n+ 1)π, 0), n an integer.
21. f(x) = x4 + cx2 = x2 x2 + c . Note that f is an even function. For c ≥ 0, the only x-intercept is the point (0, 0). We
calculate f 0(x) = 4x3 + 2cx = 4x x2 + 12c ⇒ f 00(x) = 12x2 + 2c. If c ≥ 0, x = 0 is the only critical point and there
is no inflection point. As we can see from the examples, there is no change in the basic shape of the graph for c ≥ 0; it merely
becomes steeper as c increases. For c = 0, the graph is the simple curve y = x4.For c < 0, there are x-intercepts at 0
and at±√−c. Also, there is a maximum at (0, 0), and there are
minima at ± − 12c,− 1
4c2 . As c→−∞, the x-coordinates of
these minima get larger in absolute value, and the minimum points
move downward. There are inflection points at ± − 16c,− 5
36c2 ,
which also move away from the origin as c→−∞.
23. Note that c = 0 is a transitional value at which the graph consists of the x-axis. Also, we can see that if we substitute −c for c,
the function f(x) = cx
1 + c2x2will be reflected in the x-axis, so we investigate only positive values of c (except c = −1, as a
demonstration of this reflective property). Also, f is an odd function. limx→±∞
f(x) = 0, so y = 0 is a horizontal asymptote
150 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
for all c. We calculate f 0(x) = (1 + c2x2)c− cx(2c2x)
(1 + c2x2)2= −c(c2x2 − 1)
(1 + c2x2)2. f 0(x) = 0 ⇔ c2x2 − 1 = 0 ⇔
x = ±1/c. So there is an absolute maximum value of f(1/c) = 12
and an absolute minimum value of f(−1/c) = − 12
. These
extrema
have the same value regardless of c, but the maximum points move closer to the y-axis as c increases.
f 00(x) =(−2c3x)(1 + c2x2)2 − (−c3x2 + c)[2(1 + c2x2)(2c2x)]
(1 + c2x2)4
=(−2c3x)(1 + c2x2) + (c3x2 − c)(4c2x)
(1 + c2x2)3=2c3x(c2x2 − 3)(1 + c2x2)3
f 00(x) = 0 ⇔ x = 0 or ±√3/c, so there are inflection points at (0, 0) and
at ±√3/c,±√3/4 . Again, the y-coordinate of the inflection points does not
depend on c, but as c increases, both inflection points approach the y-axis.
25. f(x) = cx+ sinx ⇒ f 0(x) = c+ cosx ⇒ f 00(x) = − sinx
f(−x) = −f(x), so f is an odd function and its graph is symmetric with respect to the origin.
f(x) = 0 ⇔ sinx = −cx, so 0 is always an x-intercept.
f 0(x) = 0 ⇔ cosx = −c, so there is no critical number when |c| > 1. If |c| ≤ 1, then there are infinitely
many critical numbers. If x1 is the unique solution of cosx = −c in the interval [0, π], then the critical numbers are 2nπ±x1,
where n ranges over the integers. (Special cases: When c = 1, x1 = 0; when c = 0, x = π2
; and when c = −1, x1 = π.)
f 00(x) < 0 ⇔ sinx > 0, so f is CD on intervals of the form (2nπ, (2n+ 1)π). f is CU on intervals of the form
((2n− 1)π, 2nπ). The inflection points of f are the points (2nπ, 2nπc), where n is an integer.
If c ≥ 1, then f 0(x) ≥ 0 for all x, so f is increasing and has no extremum. If c ≤ −1, then f 0(x) ≤ 0 for all x, so f is
decreasing and has no extremum. If |c| < 1, then f 0(x) > 0 ⇔ cosx > −c ⇔ x is in an interval of the form
(2nπ − x1, 2nπ + x1) for some integer n. These are the intervals on which f is increasing. Similarly, we
find that f is decreasing on the intervals of the form (2nπ + x1, 2(n+ 1)π − x1). Thus, f has local maxima at the points
2nπ + x1, where f has the values c(2nπ + x1) + sinx1 = c(2nπ + x1) +√1− c2, and f has local minima at the points
2nπ − x1, where we have f(2nπ − x1) = c(2nπ − x1)− sinx1 = c(2nπ − x1)−√1− c2.
The transitional values of c are −1 and 1. The inflection points move vertically, but not horizontally, when c changes.
When |c| ≥ 1, there is no extremum. For |c| < 1, the maxima are spaced
2π apart horizontally, as are the minima. The horizontal spacing between
maxima and adjacent minima is regular (and equals π) when c = 0, but
the horizontal space between a local maximum and the nearest local
minimum shrinks as |c| approaches 1.
SECTION 4.7 OPTIMIZATION PROBLEMS ¤ 151
27. (a) f(x) = cx4 − 2x2 + 1. For c = 0, f(x) = −2x2 + 1, a parabola whose vertex, (0, 1), is the absolute maximum. For
c > 0, f(x) = cx4 − 2x2 + 1 opens upward with two minimum points. As c→ 0, the minimum points spread apart and
move downward; they are below the x-axis for 0 < c < 1 and above for c > 1. For c < 0, the graph opens downward, and
has an absolute maximum at x = 0 and no local minimum.
(b) f 0(x) = 4cx3 − 4x = 4cx(x2 − 1/c) [c 6= 0]. If c ≤ 0, 0 is the only critical number.
f 00(x) = 12cx2 − 4, so f 00(0) = −4 and there is a local maximum at
(0, f(0)) = (0, 1), which lies on y = 1− x2. If c > 0, the critical
numbers are 0 and ±1/√c. As before, there is a local maximum at
(0, f(0)) = (0, 1), which lies on y = 1− x2.
f 00 ±1/√c = 12− 4 = 8 > 0, so there is a local minimum at
x = ±1/√c. Here f ±1/√c = c(1/c2)− 2/c+ 1 = −1/c+ 1.
But ±1/√c,−1/c+ 1 lies on y = 1− x2 since 1− ±1/
√c
2
= 1− 1/c.
4.7 Optimization Problems
1. (a)First Number Second Number Product
1 22 22
2 21 42
3 20 60
4 19 76
5 18 90
6 17 102
7 16 112
8 15 120
9 14 126
10 13 130
11 12 132
We needn’t consider pairs where the first number
is larger than the second, since we can just
interchange the numbers in such cases. The
answer appears to be 11 and 12, but we have
considered only integers in the table.
(b) Call the two numbers x and y. Then x+ y = 23, so y = 23− x. Call the product P . Then
P = xy = x(23− x) = 23x− x2, so we wish to maximize the function P (x) = 23x− x2. Since P 0(x) = 23− 2x,
we see that P 0(x) = 0 ⇔ x = 232 = 11.5. Thus, the maximum value of P is P (11.5) = (11.5)2 = 132.25 and it
occurs when x = y = 11.5.
Or: Note that P 00(x) = −2 < 0 for all x, so P is everywhere concave downward and the local maximum at x = 11.5
must be an absolute maximum.
3. The two numbers are x and 100x
, where x > 0. Minimize f(x) = x+100
x. f 0(x) = 1− 100
x2=
x2 − 100x2
. The critical
number is x = 10. Since f 0(x) < 0 for 0 < x < 10 and f 0(x) > 0 for x > 10, there is an absolute minimum at x = 10.
The numbers are 10 and 10.
152 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
5. If the rectangle has dimensions x and y, then its perimeter is 2x+ 2y = 100 m, so y = 50− x. Thus, the area is
A = xy = x(50− x). We wish to maximize the function A(x) = x(50− x) = 50x− x2, where 0 < x < 50. Since
A0(x) = 50− 2x = −2(x− 25), A0(x) > 0 for 0 < x < 25 and A0(x) < 0 for 25 < x < 50. Thus, A has an absolute
maximum at x = 25, and A(25) = 252 = 625 m2. The dimensions of the rectangle that maximize its area are x = y = 25 m.
(The rectangle is a square.)
7. We need to maximize Y for N ≥ 0. Y (N) =kN
1 +N2⇒
Y 0(N) =(1 +N2)k − kN(2N)
(1 +N2)2=
k(1−N2)
(1 +N2)2=
k(1 +N)(1−N)
(1 +N2)2. Y 0(N) > 0 for 0 < N < 1 and Y 0(N) < 0
for N > 1. Thus, Y has an absolute maximum of Y (1) = 12k at N = 1.
9. (a)
The areas of the three figures are 12,500, 12,500, and 9000 ft2. There appears to be a maximum area of at least 12,500 ft2.
(b) Let x denote the length of each of two sides and three dividers.
Let y denote the length of the other two sides.
(c) Area A = length×width = y · x(d) Length of fencing = 750 ⇒ 5x+ 2y = 750
(e) 5x+ 2y = 750 ⇒ y = 375− 52x ⇒ A(x) = 375− 5
2x x = 375x− 5
2x2
(f ) A0(x) = 375− 5x = 0 ⇒ x = 75. Since A00(x) = −5 < 0 there is an absolute maximum when x = 75. Then
y = 3752 = 187.5. The largest area is 75 375
2= 14,062.5 ft2. These values of x and y are between the values in the first
and second figures in part (a). Our original estimate was low.
11. xy = 1.5× 106, so y = 1.5× 106/x. Minimize the amount of fencing, which is
3x+ 2y = 3x+ 2(1.5× 106/x) = 3x+ 3× 106/x = F (x).
F 0(x) = 3− 3× 106/x2 = 3(x2 − 106)/x2. The critical number is x = 103 and
F 0(x) < 0 for 0 < x < 103 and F 0(x) > 0 if x > 103, so the absolute minimum
occurs when x = 103 and y = 1.5× 103.
The field should be 1000 feet by 1500 feet with the middle fence parallel to the short side of the field.
13. Let b be the length of the base of the box and h the height. The surface area is 1200 = b2 +4hb ⇒ h = (1200− b2)/(4b).
The volume is V = b2h = b2(1200− b2)/4b = 300b− b3/4 ⇒ V 0(b) = 300− 34b2.
V 0(b) = 0 ⇒ 300 = 34b2 ⇒ b2 = 400 ⇒ b =
√400 = 20. Since V 0(b) > 0 for 0 < b < 20 and V 0(b) < 0 for
b > 20, there is an absolute maximum when b = 20 by the First Derivative Test for Absolute Extreme Values (see page 259).
If b = 20, then h = (1200− 202)/(4 · 20) = 10, so the largest possible volume is b2h = (20)2(10) = 4000 cm3.
SECTION 4.7 OPTIMIZATION PROBLEMS ¤ 153
15. 10 = (2w)(w)h = 2w2h, so h = 5/w2. The cost is
C(w) = 10(2w2) + 6[2(2wh) + 2hw] + 6(2w2)
= 32w2 + 36wh = 32w2 + 180/w
C0(w) = 64w − 180/w2 = 4(16w3 − 45)/w2 ⇒ w = 3 4516
is the critical number. C0(w) < 0 for 0 < w < 3 4516
and
C0(w) > 0 for w > 3 4516
. The minimum cost is C 3 4516
= 32(2.8125)2/3 + 180√2.8125 ≈ $191.28.
17. The distance from a point (x, y) on the line y = 4x+ 7 to the origin is (x− 0)2 + (y − 0)2 = x2 + y2. However, it is
easier to work with the square of the distance; that is, D(x) = x2 + y22
= x2 + y2 = x2 + (4x+ 7)2. Because the
distance is positive, its minimum value will occur at the same point as the minimum value of D.
D0(x) = 2x+ 2(4x+ 7)(4) = 34x+ 56, so D0(x) = 0 ⇔ x = − 2817
.
D00(x) = 34 > 0, so D is concave upward for all x. Thus, D has an absolute minimum at x = − 2817 . The point closest to the
origin is (x, y) = − 2817 , 4 − 28
17+ 7 = − 28
17 ,717
.
19. From the figure, we see that there are two points that are farthest away from
A(1, 0). The distance d from A to an arbitrary point P (x, y) on the ellipse is
d = (x− 1)2 + (y − 0)2 and the square of the distance is
S = d 2 = x2 − 2x+ 1 + y2 = x2 − 2x+ 1 + (4− 4x2) = −3x2 − 2x+ 5.
S0 = −6x− 2 and S0 = 0 ⇒ x = − 13 . Now S00 = −6 < 0, so we know
that S has a maximum at x = − 13 . Since −1 ≤ x ≤ 1, S(−1) = 4,
S − 13= 16
3, and S(1) = 0, we see that the maximum distance is 16
3. The corresponding y-values are
y = ± 4− 4 − 13
2= ± 32
9 = ± 43
√2 ≈ ±1.89. The points are − 1
3 ,± 43
√2 .
21. The area of the rectangle is (2x)(2y) = 4xy. Also r2 = x2 + y2 so
y =√r2 − x2, so the area is A(x) = 4x
√r2 − x2. Now
A0(x) = 4√r2 − x2 − x2√
r2 − x2= 4
r2 − 2x2√r2 − x2
. The critical number is
x = 1√2r. Clearly this gives a maximum.
y = r2 − 1√2r2
= 12r2 = 1√
2r = x, which tells us that the rectangle is a square. The dimensions are 2x =
√2 r
and 2y =√2 r.
154 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
23. The height h of the equilateral triangle with sides of length L is√32 L,
since h2 + (L/2)2 = L2 ⇒ h2 = L2 − 14L2 = 3
4L2 ⇒
h =√32 L. Using similar triangles,
√32L− y
x=
√32L
L/2=√3 ⇒
√3x =
√3
2L− y ⇒ y =
√3
2L−√3x ⇒ y =
√3
2(L− 2x).
The area of the inscribed rectangle is A(x) = (2x)y =√3x(L− 2x) = √3Lx− 2√3x2, where 0 ≤ x ≤ L/2. Now
0 = A0(x) =√3L− 4√3x ⇒ x =
√3L 4
√3 = L/4. Since A(0) = A(L/2) = 0, the maximum occurs when
x = L/4, and y =√32 L−
√34 L =
√34 L, so the dimensions are L/2 and
√34 L.
25. The area of the triangle is
A(x) = 12(2t)(r + x) = t(r + x) =
√r2 − x2(r + x). Then
0 = A0(x) = r−2x
2√r2 − x2
+√r2 − x2 + x
−2x2√r2 − x2
= − x2 + rx√r2 − x2
+√r2 − x2 ⇒
x2 + rx√r2 − x2
=√r2 − x2 ⇒ x2 + rx = r2 − x2 ⇒ 0 = 2x2 + rx− r2 = (2x− r)(x+ r) ⇒
x = 12r or x = −r. Now A(r) = 0 = A(−r) ⇒ the maximum occurs where x = 1
2r, so the triangle has height
r + 12r = 3
2r and base 2 r2 − 1
2r2= 2 3
4r2 =
√3 r.
27. The cylinder has volume V = πy2(2x). Also x2 + y2 = r2 ⇒ y2 = r2 − x2, so
V (x) = π(r2 − x2)(2x) = 2π(r2x− x3), where 0 ≤ x ≤ r.
V 0(x) = 2π r2 − 3x2 = 0 ⇒ x = r/√3. Now V (0) = V (r) = 0, so there is a
maximum when x = r/√3 and V r/
√3 = π(r2 − r2/3) 2r/
√3 = 4πr3/ 3
√3 .
29. The cylinder has surface area
2(area of the base) + (lateral surface area) = 2π(radius)2 + 2π(radius)(height)
= 2πy2 + 2πy(2x)
Now x2 + y2 = r2 ⇒ y2 = r2 − x2 ⇒ y =√r2 − x2, so the surface area is
S(x) = 2π(r2 − x2) + 4πx√r2 − x2, 0 ≤ x ≤ r
= 2πr2 − 2πx2 + 4π x√r2 − x2
SECTION 4.7 OPTIMIZATION PROBLEMS ¤ 155
Thus, S0(x) = 0− 4πx+ 4π x · 12 (r2 − x2)−1/2(−2x) + (r2 − x2)1/2 · 1
= 4π −x− x2√r2 − x2
+√r2 − x2 = 4π · −x
√r2 − x2 − x2 + r2 − x2√
r2 − x2
S0(x) = 0 ⇒ x√r2 − x2 = r2 − 2x2 ( ) ⇒ x
√r2 − x2
2= (r2 − 2x2)2 ⇒
x2(r2 − x2) = r4 − 4r2x2 + 4x4 ⇒ r2x2 − x4 = r4 − 4r2x2 + 4x4 ⇒ 5x4 − 5r2x2 + r4 = 0.
This is a quadratic equation in x2. By the quadratic formula, x2 = 5±√510 r2, but we reject the root with the + sign since it
doesn’t satisfy ( ). [The right side is negative and the left side is positive.] So x = 5−√510
r. Since S(0) = S(r) = 0, the
maximum surface area occurs at the critical number and x2 = 5−√510
r2 ⇒ y2 = r2 − 5−√510
r2 = 5+√5
10r2 ⇒
the surface area is
2π 5+√5
10r2 + 4π 5−√5
105+
√5
10r2 = πr2 2 · 5+
√5
10+ 4
(5−√5)(5+
√5)
10 = πr2 5+√5
5+ 2
√205
= πr2 5+√5+ 2·2√55
= πr2 5+ 5√5
5= πr2 1 +
√5 .
31. xy = 384 ⇒ y = 384/x. Total area is
A(x) = (8 + x)(12 + 384/x) = 12(40 + x+ 256/x), so
A0(x) = 12(1− 256/x2) = 0 ⇒ x = 16. There is an absolute minimum
when x = 16 since A0(x) < 0 for 0 < x < 16 and A0(x) > 0 for x > 16.
When x = 16, y = 384/16 = 24, so the dimensions are 24 cm and 36 cm.
33. Let x be the length of the wire used for the square. The total area is
A(x) =x
4
2
+1
2
10− x
3
√3
2
10− x
3
= 116x2 +
√3
36(10− x)2, 0 ≤ x ≤ 10
A0(x) = 18x−
√3
18 (10− x) = 0 ⇔ 972x+
4√3
72 x− 40√3
72 = 0 ⇔ x = 40√3
9+ 4√3
. Now A(0) =√3
36100 ≈ 4.81,
A(10) = 10016= 6.25 and A 40
√3
9+ 4√3≈ 2.72, so
(a ) The maximum area occurs when x = 10 m, and all the wire is used for the square.
(b) The minimum area occurs when x = 40√3
9+ 4√3≈ 4.35 m.
35. The volume is V = πr2h and the surface area is
S(r) = πr2 + 2πrh = πr2 + 2πrV
πr2= πr2 +
2V
r.
S0(r) = 2πr − 2V
r2= 0 ⇒ 2πr3 = 2V ⇒ r = 3 V
πcm.
This gives an absolute minimum since S0(r) < 0 for 0 < r < 3 V
πand S0(r) > 0 for r > 3 V
π.
When r = 3 V
π, h = V
πr2=
V
π(V/π)2/3= 3 V
πcm.
156 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
37. h2 + r2 = R2 ⇒ V = π3 r
2h = π3 (R
2 − h2)h = π3 (R
2h− h3).
V 0(h) = π3(R2 − 3h2) = 0 when h = 1√
3R. This gives an absolute maximum, since
V 0(h) > 0 for 0 < h < 1√3R and V 0(h) < 0 for h > 1√
3R. The maximum volume is
V 1√3R = π
31√3R3 − 1
3√3R3 = 2
9√3πR3.
39. By similar triangles, HR=
H − h
r(1). The volume of the inner cone is V = 1
3πr2h,
so we’ll solve (1) for h. Hr
R= H − h ⇒
h = H − Hr
R=
HR−Hr
R=
H
R(R− r) (2).
Thus, V(r) = π
3r2 · H
R(R− r) =
πH
3R(Rr2 − r3) ⇒
V 0(r) =πH
3R(2Rr − 3r2) = πH
3Rr(2R− 3r).
V 0(r) = 0 ⇒ r = 0 or 2R = 3r ⇒ r = 23R and from (2), h = H
RR− 2
3R =H
R13R = 1
3H.
V 0(r) changes from positive to negative at r = 23R, so the inner cone has a maximum volume of
V = 13πr2h = 1
3π 2
3R
2 13H = 4
27· 13πR2H, which is approximately 15% of the volume of the larger cone.
41. P (R) = E2R
(R+ r)2⇒
P 0(R) =(R+ r)2 ·E2 −E2R · 2(R+ r)
[(R+ r)2]2=(R2 + 2Rr + r2)E2 − 2E2R2 − 2E2Rr
(R+ r)4
=E2r2 −E2R2
(R+ r)4=
E2(r2 −R2)
(R+ r)4=
E2(r +R)(r −R)
(R+ r)4=
E2(r −R)
(R+ r)3
P 0(R) = 0 ⇒ R = r ⇒ P (r) =E2r
(r + r)2=
E2r
4r2=
E2
4r.
The expression for P 0(R) shows that P 0(R) > 0 for R < r and P 0(R) < 0 for R > r. Thus, the maximum value of the
power is E2/(4r), and this occurs when R = r.
43. S = 6sh− 32s
2 cot θ + 3s2√32 csc θ
(a) dS
dθ= 3
2s2 csc2 θ − 3s2
√32csc θ cot θ or 3
2s2 csc θ csc θ −√3 cot θ .
(b) dS
dθ= 0 when csc θ −√3 cot θ = 0 ⇒ 1
sin θ−√3 cos θ
sin θ= 0 ⇒ cos θ = 1√
3. The First Derivative Test shows
that the minimum surface area occurs when θ = cos−1 1√3≈ 55◦.
SECTION 4.7 OPTIMIZATION PROBLEMS ¤ 157
(c) If cos θ = 1√3
, then cot θ = 1√2
and csc θ =√3√2
, so the surface area is
S = 6sh− 32s2 1√
2+ 3s2
√32
√3√2= 6sh− 3
2√2s2 + 9
2√2s2
= 6sh+ 6
2√2s2 = 6s h+ 1
2√2s
45. Here T (x) =√x2 + 25
6+5− x
8, 0 ≤ x ≤ 5 ⇒ T 0(x) =
x
6√x2 + 25
− 1
8= 0 ⇔ 8x = 6
√x2 + 25 ⇔
16x2 = 9(x2 + 25) ⇔ x = 15√7
. But 15√7> 5, so T has no critical number. Since T (0) ≈ 1.46 and T (5) ≈ 1.18, he
should row directly to B.
47. There are (6− x) km over land and√x2 + 4 km under the river.
We need to minimize the cost C (measured in $100,000) of the pipeline.
C(x) = (6− x)(4) +√x2 + 4 (8) ⇒
C0(x) = −4 + 8 · 12(x2 + 4)−1/2(2x) = −4 + 8x√
x2 + 4.
C0(x) = 0 ⇒ 4 =8x√x2 + 4
⇒ √x2 + 4 = 2x ⇒ x2 + 4 = 4x2 ⇒ 4 = 3x2 ⇒ x2 = 4
3⇒
x = 2/√3 [0 ≤ x ≤ 6]. Compare the costs for x = 0, 2/
√3, and 6. C(0) = 24 + 16 = 40,
C 2/√3 = 24− 8/√3 + 32/√3 = 24 + 24/√3 ≈ 37.9, and C(6) = 0 + 8
√40 ≈ 50.6. So the minimum cost is about
$3.79 million when P is 6− 2/√3 ≈ 4.85 km east of the refinery.
49. The total illumination is I(x) = 3k
x2+
k
(10− x)2, 0 < x < 10. Then
I0(x) =−6kx3
+2k
(10− x)3= 0 ⇒ 6k(10− x)3 = 2kx3 ⇒
3(10− x)3 = x3 ⇒ 3√3 (10− x) = x ⇒ 10 3
√3− 3
√3x = x ⇒ 10 3
√3 = x+ 3
√3x ⇒
10 3√3 = 1 + 3
√3 x ⇒ x =
10 3√3
1 + 3√3≈ 5.9 ft. This gives a minimum since I 00(x) > 0 for 0 < x < 10.
51. Every line segment in the first quadrant passing through (a, b) with endpoints on the x-
and y-axes satisfies an equation of the form y − b = m(x− a), where m < 0. By setting
x = 0 and then y = 0, we find its endpoints, A(0, b− am) and B a− bm, 0 . The
distance d from A to B is given by d = [ a− bm− 0]2 + [0− (b− am)]2.
It follows that the square of the length of the line segment, as a function of m, is given by
S(m) = a− b
m
2
+ (am− b)2 = a2 − 2ab
m+
b2
m2+ a2m2 − 2abm+ b2. Thus,
158 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
S0(m) =2ab
m2− 2b2
m3+ 2a2m− 2ab = 2
m3(abm− b2 + a2m4 − abm3)
=2
m3[b(am− b) + am3(am− b)] =
2
m3(am− b)(b+ am3)
Thus, S0(m) = 0 ⇔ m = b/a or m = − 3 ba . Since b/a > 0 and m < 0, m must equal− 3 b
a . Since 2
m3< 0, we see
that S0(m) < 0 for m < − 3 ba
and S0(m) > 0 for m > − 3 ba
. Thus, S has its absolute minimum value when m = − 3 ba
.
That value is
S − 3 ba= a+ b 3 a
b
2
+ −a 3 ba− b
2
= a+3√ab2
2
+3√a2b+ b
2
= a2 + 2a4/3b2/3 + a2/3b4/3 + a4/3b2/3 + 2a2/3b4/3 + b2 = a2 + 3a4/3b2/3 + 3a2/3b4/3 + b2
The last expression is of the form x3 + 3x2y + 3xy2 + y3 [= (x+ y)3] with x = a2/3 and y = b2/3,
so we can write it as (a2/3 + b2/3)3 and the shortest such line segment has length√S = (a2/3 + b2/3)3/2.
53. (a) If c(x) = C(x)
x, then, by Quotient Rule, we have c0(x) = xC0(x)− C(x)
x2. Now c0(x) = 0 when xC0(x)− C(x) = 0
and this gives C0(x) =C(x)
x= c(x). Therefore, the marginal cost equals the average cost.
(b) (i) C(x) = 16,000 + 200x+ 4x3/2, C(1000) = 16,000 + 200,000 + 40,000√10 ≈ 216,000 + 126,491, so
C(1000) ≈ $342,491. c(x) = C(x)/x =16,000x
+ 200 + 4x1/2, c(1000) ≈ $342.49/unit. C0(x) = 200 + 6x1/2,
C0(1000) = 200 + 60√10 ≈ $389.74/unit.
(ii) We must have C0(x) = c(x) ⇔ 200 + 6x1/2 =16,000x
+ 200 + 4x1/2 ⇔ 2x3/2 = 16,000 ⇔
x = (8,000)2/3 = 400 units. To check that this is a minimum, we calculate
c0(x) =−16,000
x2+
2√x=2
x2(x3/2 − 8000). This is negative for x < (8000)2/3 = 400, zero at x = 400,
and positive for x > 400, so c is decreasing on (0, 400) and increasing on (400,∞). Thus, c has an absolute minimum
at x = 400. [Note: c00(x) is not positive for all x > 0.]
(iii) The minimum average cost is c(400) = 40 + 200 + 80 = $320/unit.
55. (a) We are given that the demand function p is linear and p(27,000) = 10, p(33,000) = 8, so the slope is
10− 827,000− 33,000 = − 1
3000and an equation of the line is y − 10 = − 1
3000(x− 27,000) ⇒
y = p(x) = − 13000
x+ 19 = 19− (x/3000).
(b) The revenue is R(x) = xp(x) = 19x− (x2/3000) ⇒ R0(x) = 19− (x/1500) = 0 when x = 28,500. Since
R00(x) = −1/1500 < 0, the maximum revenue occurs when x = 28,500 ⇒ the price is p(28,500) = $9.50.
SECTION 4.7 OPTIMIZATION PROBLEMS ¤ 159
57. (a) As in Example 6, we see that the demand function p is linear. We are given that p(1000) = 450 and deduce that
p(1100) = 440, since a $10 reduction in price increases sales by 100 per week. The slope for p is 440− 4501100− 1000 = − 1
10,
so an equation is p− 450 = − 110(x− 1000) or p(x) = − 1
10x+ 550.
(b) R(x) = xp(x) = − 110x2 + 550x. R0(x) = − 1
5x+ 550 = 0 when x = 5(550) = 2750.
p(2750) = 275, so the rebate should be 450− 275 = $175.
(c) C(x) = 68,000 + 150x ⇒ P (x) = R(x)−C(x) = − 110x2 + 550x− 68,000− 150x = − 1
10x2 + 400x− 68,000,
P 0(x) = − 15x+ 400 = 0 when x = 2000. p(2000) = 350. Therefore, the rebate to maximize profits should be
450− 350 = $100.
59. Here s2 = h2 + b2/4, so h2 = s2 − b2/4. The area is A = 12 b s2 − b2/4.
Let the perimeter be p, so 2s+ b = p or s = (p− b)/2 ⇒
A(b) = 12 b (p− b)2/4− b2/4 = b p2 − 2pb/4. Now
A0(b) =p2 − 2pb4
− bp/4
p2 − 2pb =−3pb+ p2
4 p2 − 2pb .
Therefore, A0(b) = 0 ⇒ −3pb+ p2 = 0 ⇒ b = p/3. Since A0(b) > 0 for b < p/3 and A0(b) < 0 for b > p/3, there
is an absolute maximum when b = p/3. But then 2s+ p/3 = p, so s = p/3 ⇒ s = b ⇒ the triangle is equilateral.
61. Note that |AD| = |AP |+ |PD| ⇒ 5 = x+ |PD| ⇒ |PD| = 5− x.
Using the Pythagorean Theorem for ∆PDB and ∆PDC gives us
L(x) = |AP |+ |BP |+ |CP | = x+ (5− x)2 + 22 + (5− x)2 + 32
= x+√x2 − 10x+ 29 +√x2 − 10x+ 34 ⇒
L0(x) = 1 +x− 5√
x2 − 10x+ 29 +x− 5√
x2 − 10x+ 34 . From the graphs of L
and L0, it seems that the minimum value of L is about L(3.59) = 9.35 m.
63. The total time is
T (x) = (time from A to C) + (time from C to B)
=
√a2 + x2
v1+
b2 + (d− x)2
v2, 0 < x < d
T 0(x) =x
v1√a2 + x2
− d− x
v2 b2 + (d− x)2=sin θ1v1
− sin θ2v2
The minimum occurs when T 0(x) = 0 ⇒ sin θ1v1
=sin θ2v2
.
[Note: T 00(x) > 0]
160 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
65. y2 = x2 + z2, but triangles CDE and BCA are similar, so
z/8 = x/ 4√x− 4 ⇒ z = 2x/
√x− 4. Thus, we minimize
f(x) = y2 = x2 + 4x2/(x− 4) = x3/(x− 4), 4 < x ≤ 8.
f 0(x) =(x− 4)(3x2)− x3
(x− 4)2 =x2[3(x− 4)− x]
(x− 4)2 =2x2(x− 6)(x− 4)2 = 0
when x = 6. f 0(x) < 0 when x < 6, f 0(x) > 0 when x > 6, so the minimum
occurs when x = 6 in.
67. It suffices to maximize tan θ. Now
3t
1= tan(ψ + θ) =
tanψ + tan θ
1− tanψ tan θ =t+ tan θ
1− t tan θ. So
3t(1− t tan θ) = t+ tan θ ⇒ 2t = (1 + 3t2) tan θ ⇒ tan θ =2t
1 + 3t2.
Let f(t) = tan θ = 2t
1 + 3t2⇒ f 0(t) =
2 1 + 3t2 − 2t(6t)(1 + 3t2)2
=2 1− 3t2(1 + 3t2)2
= 0 ⇔ 1− 3t2 = 0 ⇔
t = 1√3
since t ≥ 0. Now f 0(t) > 0 for 0 ≤ t < 1√3
and f 0(t) < 0 for t > 1√3
, so f has an absolute maximum when t = 1√3
and tan θ =2 1/
√3
1 + 3 1/√3
2 =1√3
⇒ θ = π6
. Substituting for t and θ in 3t = tan(ψ + θ) gives us
√3 = tan ψ + π
6⇒ ψ = π
6.
69. In the small triangle with sides a and c and hypotenuse W , sin θ = a
Wand
cos θ =c
W. In the triangle with sides b and d and hypotenuse L, sin θ = d
Land
cos θ =b
L. Thus, a =W sin θ, c =W cos θ, d = L sin θ, and b = L cos θ, so the
area of the circumscribed rectangle is
A(θ) = (a+ b)(c+ d) = (W sin θ + L cos θ)(W cos θ + L sin θ)
=W 2 sin θ cos θ +WL sin2 θ + LW cos2 θ + L2 sin θ cos θ
= LW sin2 θ + LW cos2 θ + (L2 +W 2) sin θ cos θ
= LW (sin2 θ + cos2 θ) + (L2 +W 2) · 12 · 2 sin θ cos θ = LW + 12 (L
2 +W 2) sin 2θ, 0 ≤ θ ≤ π2
This expression shows, without calculus, that the maximum value of A(θ) occurs when sin 2θ = 1 ⇔ 2θ = π2 ⇒
θ = π4
. So the maximum area is A π4= LW + 1
2(L2 +W 2) = 1
2(L2 + 2LW +W 2) = 1
2(L+W )2.
SECTION 4.8 NEWTON’S METHOD ¤ 161
71. (a) If k = energy/km over land, then energy/km over water = 1.4k.
So the total energy is E = 1.4k√25 + x2 + k(13− x), 0 ≤ x ≤ 13,
and so dE
dx=
1.4kx
(25 + x2)1/2− k.
Set dEdx
= 0: 1.4kx = k(25 + x2)1/2 ⇒ 1.96x2 = x2 + 25 ⇒ 0.96x2 = 25 ⇒ x = 5√0.96
≈ 5.1.
Testing against the value of E at the endpoints: E(0) = 1.4k(5) + 13k = 20k, E(5.1) ≈ 17.9k, E(13) ≈ 19.5k.
Thus, to minimize energy, the bird should fly to a point about 5.1 km from B.
(b) If W/L is large, the bird would fly to a point C that is closer to B than to D to minimize the energy used flying over water.
If W/L is small, the bird would fly to a point C that is closer to D than to B to minimize the distance of the flight.
E =W√25 + x2 + L(13− x) ⇒ dE
dx=
Wx√25 + x2
− L = 0 when W
L=
√25 + x2
x. By the same sort of
argument as in part (a), this ratio will give the minimal expenditure of energy if the bird heads for the point x km from B.
(c) For flight direct to D, x = 13, so from part (b), W/L =
√25+ 132
13≈ 1.07. There is no value of W/L for which the bird
should fly directly to B. But note that limx→0+
(W/L) =∞, so if the point at which E is a minimum is close to B, then
W/L is large.
(d) Assuming that the birds instinctively choose the path that minimizes the energy expenditure, we can use the equation for
dE/dx = 0 from part (a) with 1.4k = c, x = 4, and k = 1: c(4) = 1 · (25 + 42)1/2 ⇒ c =√41/4 ≈ 1.6.
4.8 Newton's Method
1. (a) The tangent line at x = 1 intersects the x-axis at x ≈ 2.3, so
x2 ≈ 2.3. The tangent line at x = 2.3 intersects the x-axis at
x ≈ 3, so x3 ≈ 3.0.
(b) x1 = 5 would not be a better first approximation than x1 = 1 since the tangent line is nearly horizontal. In fact, the second
approximation for x1 = 5 appears to be to the left of x = 1.
3. Since x1 = 3 and y = 5x− 4 is tangent to y = f(x) at x = 3, we simply need to find where the tangent line intersects the
x-axis. y = 0 ⇒ 5x2 − 4 = 0 ⇒ x2 =45
.
5. f(x) = x3 + 2x− 4 ⇒ f 0(x) = 3x2 + 2, so xn+1 = xn − x3n + 2xn − 43x2n + 2
. Now x1 = 1 ⇒
x2 = 1− 1 + 2− 43 · 12 + 2 = 1−
−15= 1.2 ⇒ x3 = 1.2− (1.2)3 + 2(1.2)− 4
3(1.2)2 + 2≈ 1.1797.
162 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
7. f(x) = x5 − x− 1 ⇒ f 0(x) = 5x4 − 1, so xn+1 = xn − x5n − xn − 15x4n − 1 . Now x1 = 1 ⇒
x2 = 1− 1− 1− 15− 1 = 1− − 1
4= 1.25 ⇒ x3 = 1.25− (1.25)5 − 1.25− 1
5(1.25)4 − 1 ≈ 1.1785.
9. f(x) = x3 + x+ 3 ⇒ f 0(x) = 3x2 + 1, so xn+1 = xn − x3n + xn + 3
3x2n + 1.
Now x1 = −1 ⇒
x2 = −1− (−1)3 + (−1) + 33(−1)2 + 1 = −1− −1− 1 + 3
3 + 1= −1− 1
4= −1.25.
Newton’s method follows the tangent line at (−1, 1) down to its intersection with
the x-axis at (−1.25, 0), giving the second approximation x2 = −1.25.
11. To approximate x = 5√20 (so that x5 = 20), we can take f(x) = x5 − 20. So f 0(x) = 5x4, and thus,
xn+1 = xn − x5n − 205x4n
. Since 5√32 = 2 and 32 is reasonably close to 20, we’ll use x1 = 2. We need to find approximations
until they agree to eight decimal places. x1 = 2 ⇒ x2 = 1.85, x3 ≈ 1.82148614, x4 ≈ 1.82056514,
x5 ≈ 1.82056420 ≈ x6. So 5√20 ≈ 1.82056420, to eight decimal places.
Here is a quick and easy method for finding the iterations for Newton’s method on a programmable calculator.
(The screens shown are from the TI-84 Plus, but the method is similar on other calculators.) Assign f(x) = x5 − 20
to Y1, and f 0(x) = 5x4 to Y2. Now store x1 = 2 in X and then enter X−Y1/Y2 → X to get x2 = 1.85. By successively
pressing the ENTER key, you get the approximations x3, x4, . . . .
In Derive, load the utility file SOLVE. Enter NEWTON(xˆ5-20,x,2) and then APPROXIMATE to get
[2, 1.85, 1.82148614, 1.82056514, 1.82056420]. You can request a specific iteration by adding a fourth argument. For
example, NEWTON(xˆ5-20,x,2,2) gives [2, 1.85, 1.82148614].
In Maple, make the assignments f := x→ xˆ5− 20;, g := x→ x− f(x)/D(f)(x);, and x := 2.;. Repeatedly execute
the command x := g(x); to generate successive approximations.
In Mathematica, make the assignments f [x_ ] := xˆ5− 20, g[x_ ] := x− f [x]/f 0[x], and x = 2. Repeatedly execute the
command x = g[x] to generate successive approximations.
13. f(x) = x4 − 2x3 + 5x2 − 6 ⇒ f 0(x) = 4x3 − 6x2 + 10x ⇒ xn+1 = xn − x4n − 2x3n + 5x2n − 64x3n − 6x2n + 10xn . We need to find
approximations until they agree to six decimal places. We’ll let x1 equal the midpoint of the given interval, [1, 2].
x1 = 1.5 ⇒ x2 = 1.2625, x3 ≈ 1.218808, x4 ≈ 1.217563, x5 ≈ 1.217562 ≈ x6. So the root is 1.217562 to six decimal
places.
SECTION 4.8 NEWTON’S METHOD ¤ 163
15. sinx = x2, so f(x) = sinx− x2 ⇒ f 0(x) = cosx− 2x ⇒
xn+1 = xn − sinxn − x2ncosxn − 2xn . From the figure, the positive root of sin x = x2 is
near 1. x1 = 1 ⇒ x2 ≈ 0.891396, x3 ≈ 0.876985, x4 ≈ 0.876726 ≈ x5. So
the positive root is 0.876726, to six decimal places.
17. From the graph, we see that there appear to be points of intersection near
x = −0.7 and x = 1.2. Solving x4 = 1 + x is the same as solving
f(x) = x4 − x− 1 = 0. f(x) = x4 − x− 1 ⇒ f 0(x) = 4x3 − 1,
so xn+1 = xn − x4n − xn − 14x3n − 1 .
x1 = −0.7 x1 = 1.2
x2 ≈ −0.725253 x2 ≈ 1.221380x3 ≈ −0.724493 x3 ≈ 1.220745x4 ≈ −0.724492 ≈ x5 x4 ≈ 1.220744 ≈ x5
To six decimal places, the roots of the equation are−0.724492 and 1.220744.
19. From the graph, we see that there appear to be points of intersection near
x = −0.5 and x = 1.5. Solving 3√x = x2 − 1 is the same as solving
f(x) =3√x− x2 + 1 = 0. f(x) =
3√x− x2 + 1 ⇒
f 0(x) = 13x−2/3 − 2x, so xn+1 = xn −
3√xn − x2n + 1
13x−2/3n − 2xn
.
x1 = −0.5x2 ≈ −0.471421x3 ≈ −0.471074 ≈ x4
x1 = 1.5
x2 ≈ .461653
x3 ≈ 1.461070 ≈ x4
To six decimal places, the roots are −0.471074 and 1.461070.
21. From the graph, there appears to be a point of intersection near x = 0.6.
Solving cosx =√x is the same as solving f(x) = cosx−
√x = 0.
f(x) = cosx−√x ⇒ f 0(x) = − sinx− 1/ 2
√x , so
xn+1 = xn − cosxn −√xn
− sinxn − 1/ 2√x
. Now x1 = 0.6 ⇒ x2 ≈ 0.641928,
x3 ≈ 0.641714 ≈ x4. To six decimal places, the root of the equation is 0.641714.
164 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
23. f(x) = x6 − x5 − 6x4 − x2 + x+ 10 ⇒f 0(x) = 6x5 − 5x4 − 24x3 − 2x+ 1 ⇒
xn+1 = xn − x6n − x5n − 6x4n − x2n + xn + 10
6x5n − 5x4n − 24x3n − 2xn + 1 .
From the graph of f , there appear to be roots near −1.9, −1.2, 1.1, and 3.
x1 = −1.9x2 ≈ −1.94278290x3 ≈ −1.93828380x4 ≈ −1.93822884x5 ≈ −1.93822883 ≈ x6
x1 = −1.2x2 ≈ −1.22006245x3 ≈ −1.21997997 ≈ x4
x1 = 1.1
x2 ≈ 1.14111662x3 ≈ 1.13929741x4 ≈ 1.13929375 ≈ x5
x1 = 3
x2 ≈ 2.99x3 ≈ 2.98984106x4 ≈ 2.98984102 ≈ x5
To eight decimal places, the roots of the equation are−1.93822883,−1.21997997, 1.13929375, and 2.98984102.
25. From the graph, y = x2√2− x− x2 and y = 1 intersect twice, at x ≈ −2 and
at x ≈ −1. f(x) = x2√2− x− x2 − 1 ⇒
f 0(x) = x2 · 12(2− x− x2)−1/2(−1− 2x) + (2− x− x2)1/2 · 2x
= 12x(2− x− x2)−1/2[x(−1− 2x) + 4(2− x− x2)]
=x(8− 5x− 6x2)2 (2 + x)(1− x)
,
so xn+1 = xn −x2n√2− xn − x2n − 1
xn(8− 5xn − 6x2n)2 (2 + xn)(1− xn)
. Trying x1 = −2 won’t work because f 0(−2) is undefined, so we’ll
try x1 = −1.95.x1 = −1.95x2 ≈ −1.98580357x3 ≈ −1.97899778x4 ≈ −1.97807848x5 ≈ −1.97806682x6 ≈ −1.97806681 ≈ x7
x1 = −0.8x2 ≈ −0.82674444x3 ≈ −0.82646236x4 ≈ −0.82646233 ≈ x5
To eight decimal places, the roots of the equation are−1.97806681 and −0.82646233.
27. (a) f(x) = x2 − a ⇒ f 0(x) = 2x, so Newton’s method gives
xn+1 = xn − x2n − a
2xn= xn − 1
2xn +
a
2xn=1
2xn +
a
2xn=1
2xn +
a
xn.
(b) Using (a) with a = 1000 and x1 =√900 = 30, we get x2 ≈ 31.666667, x3 ≈ 31.622807, and x4 ≈ 31.622777 ≈ x5.
So√1000 ≈ 31.622777.
SECTION 4.8 NEWTON’S METHOD ¤ 165
29. f(x) = x3 − 3x+ 6 ⇒ f 0(x) = 3x2 − 3. If x1 = 1, then f 0(x1) = 0 and the tangent line used for approximating x2 is
horizontal. Attempting to find x2 results in trying to divide by zero.
31. For f(x) = x1/3, f 0(x) = 13x−2/3 and
xn+1 = xn − f(xn)
f 0(xn)= xn − x
1/3n
13x−2/3n
= xn − 3xn = −2xn.
Therefore, each successive approximation becomes twice as large as the
previous one in absolute value, so the sequence of approximations fails to
converge to the root, which is 0. In the figure, we have x1 = 0.5,
x2 = −2(0.5) = −1, and x3 = −2(−1) = 2.
33. (a) f(x) = x6 − x4 + 3x3 − 2x ⇒ f 0(x) = 6x5 − 4x3 + 9x2 − 2 ⇒f 00(x) = 30x4 − 12x2 + 18x. To find the critical numbers of f , we’ll find the
zeros of f 0. From the graph of f 0, it appears there are zeros at approximately
x = −1.3, −0.4, and 0.5. Try x1 = −1.3 ⇒
x2 = x1 − f 0(x1)f 00(x1)
≈ −1.293344 ⇒ x3 ≈ −1.293227 ≈ x4.
Now try x1 = −0.4 ⇒ x2 ≈ −0.443755 ⇒ x3 ≈ −0.441735 ⇒ x4 ≈ −0.441731 ≈ x5. Finally try
x1 = 0.5 ⇒ x2 ≈ 0.507937 ⇒ x3 ≈ 0.507854 ≈ x4. Therefore, x = −1.293227, −0.441731, and 0.507854 are
all the critical numbers correct to six decimal places.
(b) There are two critical numbers where f 0 changes from negative to positive, so f changes from decreasing to increasing.
f(−1.293227) ≈ −2.0212 and f(0.507854) ≈ −0.6721, so −2.0212 is the absolute minimum value of f correct to four
decimal places.
35. y = x3 + cosx ⇒ y0 = 3x2 − sinx ⇒ y00 = 6x− cosx ⇒ y000 = 6 + sinx. Now to solve y00 = 0,
try x1 = 0, and then x2 = x1 − y00(x1)y000(x1)
=1
6⇒ x3 ≈ 0.164419 ≈ x4. For x < 0.164419, y00 < 0, and for
x > 0.164419, y00 > 0. Therefore, the point of inflection, correct to six decimal places, is
(0.164419, y(0.164419)) ≈ (0.164419, 0.990958).
37. We need to minimize the distance from (0, 0) to an arbitrary point (x, y) on the
curve y = (x− 1)2. d = x2 + y2 ⇒d(x) = x2 + [(x− 1)2]2 = x2 + (x− 1)4. When d0 = 0, d will be
minimized and equivalently, s = d2 will be minimized, so we will use Newton’s
method with f = s0 and f 0 = s00.
f(x) = 2x+ 4(x− 1)3 ⇒ f 0(x) = 2 + 12(x− 1)2, so xn+1 = xn − 2xn + 4(xn − 1)32 + 12(xn − 1)2 . Try x1 = 0.5 ⇒
x2 = 0.4, x3 ≈ 0.410127, x4 ≈ 0.410245 ≈ x5. Now d(0.410245) ≈ 0.537841 is the minimum distance and the point on
the parabola is (0.410245, 0.347810), correct to six decimal places.
166 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
39. In this case, A = 18,000, R = 375, and n = 5(12) = 60. So the formula A = R
i[1− (1 + i)−n] becomes
18,000 = 375
x[1− (1 + x)−60] ⇔ 48x = 1− (1 + x)−60 [multiply each term by (1 + x)60] ⇔
48x(1 + x)60 − (1 + x)60 + 1 = 0. Let the LHS be called f(x), so that
f 0(x) = 48x(60)(1 + x)59 + 48(1 + x)60 − 60(1 + x)59
= 12(1 + x)59[4x(60) + 4(1 + x)− 5] = 12(1 + x)59(244x− 1)
xn+1 = xn − 48xn(1 + xn)60 − (1 + xn)
60 + 1
12(1 + xn)59(244xn − 1) . An interest rate of 1% per month seems like a reasonable estimate for
x = i. So let x1 = 1% = 0.01, and we get x2 ≈ 0.0082202, x3 ≈ 0.0076802, x4 ≈ 0.0076291, x5 ≈ 0.0076286 ≈ x6.
Thus, the dealer is charging a monthly interest rate of 0.76286% (or 9.55% per year, compounded monthly).
4.9 Antiderivatives
1. f(x) = x− 3 = x1 − 3 ⇒ F (x) =x1+1
1 + 1− 3x+ C = 1
2x2 − 3x+C
Check: F 0(x) = 12(2x)− 3 + 0 = x− 3 = f(x)
3. f(x) = 12+ 3
4x2 − 4
5x3 ⇒ F (x) = 1
2x+
3
4
x2+1
2 + 1− 4
5
x3+1
3 + 1+C = 1
2x+ 1
4x3 − 1
5x4 +C
Check: F 0(x) = 12+ 1
4(3x2)− 1
5(4x3) + 0 = 1
2+ 3
4x2 − 4
5x3 = f(x)
5. f(x) = (x+ 1)(2x− 1) = 2x2 + x− 1 ⇒ F (x) = 2 13x3 + 1
2x2 − x+ C = 2
3x3 + 1
2x2 − x+ C
7. f(x) = 5x1/4 − 7x3/4 ⇒ F (x) = 5x1/4+ 1
14+ 1
− 7x3/4+ 1
34+ 1
+C = 5x5/4
5/4− 7x
7/4
7/4+ C = 4x5/4 − 4x7/4 +C
9. f(x) = 6√x− 6
√x = 6x1/2 − x1/6 ⇒
F (x) = 6x1/2+1
12+ 1
− x1/6+1
16+ 1
+ C = 6x3/2
3/2− x7/6
7/6+C = 4x3/2 − 6
7x7/6 + C
11. f(x) = 10
x9= 10x−9 has domain (−∞, 0) ∪ (0,∞), so F (x) =
⎧⎪⎪⎨⎪⎪⎩10x−8
−8 + C1 = − 5
4x8+C1 if x < 0
− 5
4x8+C2 if x > 0
See Example 1(b) for a similar problem.
13. f(u) = u4 + 3√u
u2=
u4
u2+3u1/2
u2= u2 + 3u−3/2 ⇒
F (u) =u3
3+ 3
u−3/2+1
−3/2 + 1 + C =1
3u3 + 3
u−1/2
−1/2 + C =1
3u3 − 6√
u+ C
15. g(θ) = cos θ − 5 sin θ ⇒ G(θ) = sin θ − 5(− cos θ) + C = sin θ + 5cos θ +C
SECTION 4.9 ANTIDERIVATIVES ¤ 167
17. f(t) = 2 sec t tan t+ 12 t−1/2 has domain 0, π2 and nπ − π
2 , nπ +π2
for integers n ≥ 1. The antiderivative is
F (t) = 2 sec t+ t1/2 + C0 on the interval 0, π2
or F (t) = 2 sec t+ t1/2 + Cn on the interval nπ − π2, nπ + π
2for
integers n ≥ 1.
19. f(x) = 5x4 − 2x5 ⇒ F (x) = 5 · x5
5− 2 · x
6
6+C = x5 − 1
3x6 + C.
F (0) = 4 ⇒ 05 − 13· 06 + C = 4 ⇒ C = 4, so F (x) = x5 − 1
3x6 + 4.
The graph confirms our answer since f(x) = 0 when F has a local maximum, f is
positive when F is increasing, and f is negative when F is decreasing.
21. f 00(x) = 6x+ 12x2 ⇒ f 0(x) = 6 · x2
2+ 12 · x
3
3+ C = 3x2 + 4x3 + C ⇒
f(x) = 3 · x3
3+ 4 · x
4
4+Cx+D = x3 + x4 + Cx+D [C and D are just arbitrary constants]
23. f 00(x) = 23x2/3 ⇒ f 0(x) =
2
3
x5/3
5/3+C = 2
5x5/3 + C ⇒ f(x) =
2
5
x8/3
8/3+ Cx+D = 3
20x8/3 +Cx+D
25. f 000(t) = 60t2 ⇒ f 00(t) = 20t3 + C ⇒ f 0(t) = 5t4 + Ct+D ⇒ f(t) = t5 + 12Ct
2 +Dt+E
27. f 0(x) = 1− 6x ⇒ f(x) = x− 3x2 +C. f(0) = C and f(0) = 8 ⇒ C = 8, so f(x) = x− 3x2 + 8.
29. f 0(x) =√x(6 + 5x) = 6x1/2 + 5x3/2 ⇒ f(x) = 4x3/2 + 2x5/2 + C.
f(1) = 6 +C and f(1) = 10 ⇒ C = 4, so f(x) = 4x3/2 + 2x5/2 + 4.
31. f 0(t) = 2 cos t+ sec2 t ⇒ f(t) = 2 sin t+ tan t+ C because −π/2 < t < π/2.
f π3= 2
√3/2 +
√3 +C = 2
√3 +C and f π
3= 4 ⇒ C = 4− 2√3, so f(t) = 2 sin t+ tan t+ 4− 2√3.
33. f 00(x) = 24x2 + 2x+ 10 ⇒ f 0(x) = 8x3 + x2 + 10x+ C. f 0(1) = 8 + 1 + 10 + C and f 0(1) = −3 ⇒
19 + C = −3 ⇒ C = −22, so f 0(x) = 8x3 + x2 + 10x− 22 and hence, f(x) = 2x4 + 13x3 + 5x2 − 22x+D.
f(1) = 2 + 13 + 5− 22 +D and f(1) = 5 ⇒ D = 22− 7
3 =593 , so f(x) = 2x4 + 1
3x3 + 5x2 − 22x+ 59
3 .
35. f 00(θ) = sin θ + cos θ ⇒ f 0(θ) = − cos θ + sin θ + C. f 0(0) = −1 + C and f 0(0) = 4 ⇒ C = 5, so
f 0(θ) = − cos θ + sin θ + 5 and hence, f(θ) = − sin θ − cos θ + 5θ +D. f(0) = −1 +D and f(0) = 3 ⇒ D = 4,
so f(θ) = − sin θ − cos θ + 5θ + 4.
37. f 00(x) = 2− 12x ⇒ f 0(x) = 2x− 6x2 +C ⇒ f(x) = x2 − 2x3 + Cx+D.
f(0) = D and f(0) = 9 ⇒ D = 9. f(2) = 4− 16 + 2C + 9 = 2C − 3 and f(2) = 15 ⇒ 2C = 18 ⇒
C = 9, so f(x) = x2 − 2x3 + 9x+ 9.
168 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
39. f 00(x) = 2 + cosx ⇒ f 0(x) = 2x+ sinx+ C ⇒ f(x) = x2 − cosx+ Cx+D.
f(0) = −1 +D and f(0) = −1 ⇒ D = 0. f π2= π2/4 + π
2C and f π
2= 0 ⇒ π
2C = −π2/4 ⇒
C = −π2
, so f(x) = x2 − cosx− π2x.
41. Given f 0(x) = 2x+ 1, we have f(x) = x2 + x+C. Since f passes through (1, 6), f(1) = 6 ⇒ 12 +1+C = 6 ⇒
C = 4. Therefore, f(x) = x2 + x+ 4 and f(2) = 22 + 2 + 4 = 10.
43. b is the antiderivative of f . For small x, f is negative, so the graph of its antiderivative must be decreasing. But both a and c
are increasing for small x, so only b can be f ’s antiderivative. Also, f is positive where b is increasing, which supports our
conclusion.
45. The graph of F must start at (0, 1). Where the given graph, y = f(x), has a
local minimum or maximum, the graph of F will have an inflection point.
Where f is negative (positive), F is decreasing (increasing).
Where f changes from negative to positive, F will have a minimum.
Where f changes from positive to negative, F will have a maximum.
Where f is decreasing (increasing), F is concave downward (upward).
47.
f 0(x) =
⎧⎪⎨⎪⎩2 if 0 ≤ x < 1
1 if 1 < x < 2
−1 if 2 < x ≤ 3⇒ f(x) =
⎧⎪⎨⎪⎩2x+C if 0 ≤ x < 1
x+D if 1 < x < 2
−x+E if 2 < x ≤ 3
f(0) = −1 ⇒ 2(0) +C = −1 ⇒ C = −1. Starting at the point
(0,−1) and moving to the right on a line with slope 2 gets us to the point (1, 1).
The slope for 1 < x < 2 is 1, so we get to the point (2, 2). Here we have used the fact that f is continuous. We can include the
point x = 1 on either the first or the second part of f . The line connecting (1, 1) to (2, 2) is y = x, so D = 0. The slope for
2 < x ≤ 3 is −1, so we get to (3, 1). f(3) = 1 ⇒ −3 +E = 1 ⇒ E = 4. Thus
f(x) =
⎧⎪⎨⎪⎩2x− 1 if 0 ≤ x ≤ 1x if 1 < x < 2
−x+ 4 if 2 ≤ x ≤ 3
Note that f 0(x) does not exist at x = 1 or at x = 2.
SECTION 4.9 ANTIDERIVATIVES ¤ 169
49. f(x) = sinx
1 + x2, −2π ≤ x ≤ 2π
Note that the graph of f is one of an odd function, so the graph of F will
be one of an even function.
51. v(t) = s0(t) = sin t− cos t ⇒ s(t) = − cos t− sin t+ C. s(0) = −1 + C and s(0) = 0 ⇒ C = 1, so
s(t) = − cos t− sin t+ 1.
53. a(t) = v0(t) = t− 2 ⇒ v(t) = 12 t2 − 2t+ C. v(0) = C and v(0) = 3 ⇒ C = 3, so v(t) = 1
2 t2 − 2t+ 3 and
s(t) = 16t3 − t2 + 3t+D. s(0) = D and s(0) = 1 ⇒ D = 1, and s(t) = 1
6t3 − t2 + 3t+ 1.
55. a(t) = v0(t) = 10 sin t+ 3cos t ⇒ v(t) = −10 cos t+ 3 sin t+ C ⇒ s(t) = −10 sin t− 3 cos t+ Ct+D.
s(0) = −3 +D = 0 and s(2π) = −3 + 2πC +D = 12 ⇒ D = 3 and C = 6π . Thus,
s(t) = −10 sin t− 3 cos t+ 6πt+ 3.
57. (a) We first observe that since the stone is dropped 450 m above the ground, v(0) = 0 and s(0) = 450.
v0(t) = a(t) = −9.8 ⇒ v(t) = −9.8t+ C. Now v(0) = 0 ⇒ C = 0, so v(t) = −9.8t ⇒s(t) = −4.9t2 +D. Last, s(0) = 450 ⇒ D = 450 ⇒ s(t) = 450− 4.9t2.
(b) The stone reaches the ground when s(t) = 0. 450− 4.9t2 = 0 ⇒ t2 = 450/4.9 ⇒ t1 = 450/4.9 ≈ 9.58 s.
(c) The velocity with which the stone strikes the ground is v(t1) = −9.8 450/4.9 ≈ −93.9 m/s.
(d) This is just reworking parts (a) and (b) with v(0) = −5. Using v(t) = −9.8t+C, v(0) = −5 ⇒ 0 + C = −5 ⇒v(t) = −9.8t− 5. So s(t) = −4.9t2 − 5t+D and s(0) = 450 ⇒ D = 450 ⇒ s(t) = −4.9t2 − 5t+ 450.
Solving s(t) = 0 by using the quadratic formula gives us t = 5±√8845 (−9.8) ⇒ t1 ≈ 9.09 s.
59. By Exercise 58 with a = −9.8, s(t) = −4.9t2 + v0t+ s0 and v(t) = s0 (t) = −9.8t+ v0. So
[v(t)]2 = (−9.8t+ v0)2 = (9.8)2 t2 − 19.6v0t+ v20 = v20 + 96.04t
2 − 19.6v0t = v20 − 19.6 −4.9t2 + v0t .
But −4.9t2 + v0t is just s(t) without the s0 term; that is, s(t)− s0. Thus, [v(t)]2 = v20 − 19.6 [s(t)− s0].
61. Using Exercise 58 with a = −32, v0 = 0, and s0 = h (the height of the cliff ), we know that the height at time t is
s(t) = −16t2 + h. v(t) = s0(t) = −32t and v(t) = −120 ⇒ −32t = −120 ⇒ t = 3.75, so
0 = s(3.75) = −16(3.75)2 + h ⇒ h = 16(3.75)2 = 225 ft.
170 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
63. Marginal cost = 1.92− 0.002x = C0(x) ⇒ C(x) = 1.92x− 0.001x2 +K. But C(1) = 1.92− 0.001 +K = 562 ⇒
K = 560.081. Therefore, C(x) = 1.92x− 0.001x2 + 560.081 ⇒ C(100) = 742.081, so the cost of producing
100 items is $742.08.
65. Taking the upward direction to be positive we have that for 0 ≤ t ≤ 10 (using the subscript 1 to refer to 0 ≤ t ≤ 10),
a1(t) = − (9− 0.9t) = v01(t) ⇒ v1(t) = −9t+ 0.45t2 + v0, but v1(0) = v0 = −10 ⇒
v1(t) = −9t+ 0.45t2 − 10 = s01(t) ⇒ s1(t) = − 92t2 + 0.15t3 − 10t+ s0. But s1(0) = 500 = s0 ⇒
s1(t) = − 92 t2 + 0.15t3 − 10t+ 500. s1(10) = −450 + 150− 100 + 500 = 100, so it takes
more than 10 seconds for the raindrop to fall. Now for t > 10, a(t) = 0 = v0(t) ⇒
v(t) = constant = v1(10) = −9(10) + 0.45(10)2 − 10 = −55 ⇒ v(t) = −55.
At 55 m/s, it will take 100/55 ≈ 1.8 s to fall the last 100 m. Hence, the total time is 10 + 10055= 130
11≈ 11.8 s.
67. a(t) = k, the initial velocity is 30 mi/h = 30 · 52803600
= 44 ft/s, and the final velocity (after 5 seconds) is
50 mi/h = 50 · 52803600 =
2203 ft/s. So v(t) = kt+ C and v(0) = 44 ⇒ C = 44. Thus, v(t) = kt+ 44 ⇒
v(5) = 5k + 44. But v(5) = 2203
, so 5k + 44 = 2203
⇒ 5k = 883
⇒ k = 8815≈ 5.87 ft/s2.
69. Let the acceleration be a(t) = k km/h2. We have v(0) = 100 km/h and we can take the initial position s(0) to be 0.
We want the time tf for which v(t) = 0 to satisfy s(t) < 0.08 km. In general, v0(t) = a(t) = k, so v(t) = kt+ C,
where C = v(0) = 100. Now s0(t) = v(t) = kt+ 100, so s(t) = 12kt2 + 100t+D, where D = s(0) = 0.
Thus, s(t) = 12kt2 + 100t. Since v(tf ) = 0, we have ktf + 100 = 0 or tf = −100/k, so
s(tf ) =1
2k −100
k
2
+ 100 −100k
= 10,000 1
2k− 1
k= −5,000
k. The condition s(tf ) must satisfy is
−5,000k
< 0.08 ⇒ −5,0000.08
> k [k is negative] ⇒ k < −62,500 km/h2, or equivalently,
k < − 3125648
≈ −4.82 m/s2.
71. (a) First note that 90 mi/h = 90× 52803600
ft/s = 132 ft/s. Then a(t) = 4 ft/s2 ⇒ v(t) = 4t+ C, but v(0) = 0 ⇒
C = 0. Now 4t = 132 when t = 1324 = 33 s, so it takes 33 s to reach 132 ft/s. Therefore, taking s(0) = 0, we have
s(t) = 2t2, 0 ≤ t ≤ 33. So s(33) = 2178 ft. 15 minutes = 15(60) = 900 s, so for 33 < t ≤ 933 we have
v(t) = 132 ft/s ⇒ s(933) = 132(900) + 2178 = 120,978 ft = 22.9125 mi.
(b) As in part (a), the train accelerates for 33 s and travels 2178 ft while doing so. Similarly, it decelerates for 33 s and travels
2178 ft at the end of its trip. During the remaining 900− 66 = 834 s it travels at 132 ft/s, so the distance traveled is
132 · 834 = 110,088 ft. Thus, the total distance is 2178 + 110,088 + 2178 = 114,444 ft = 21.675 mi.
CHAPTER 4 REVIEW ¤ 171
(c) 45 mi = 45(5280) = 237,600 ft. Subtract 2(2178) to take care of the speeding up and slowing down, and we have
233,244 ft at 132 ft/s for a trip of 233,244/132 = 1767 s at 90 mi/h. The total time is
1767 + 2(33) = 1833 s = 30 min 33 s = 30.55 min.
(d) 37.5(60) = 2250 s. 2250− 2(33) = 2184 s at maximum speed. 2184(132) + 2(2178) = 292,644 total feet or
292,644/5280 = 55.425 mi.
4 Review
1. A function f has an absolute maximum at x = c if f(c) is the largest function value on the entire domain of f , whereas f has
a local maximum at c if f(c) is the largest function value when x is near c. See Figure 4 in Section 4.1.
2. (a) See Theorem 4.1.3.
(b) See the Closed Interval Method before Example 8 in Section 4.1.
3. (a) See Theorem 4.1.4.
(b) See Definition 4.1.6.
4. (a) See Rolle’s Theorem at the beginning of Section 4.2.
(b) See the Mean Value Theorem in Section 4.2. Geometric interpretation—there is some point P on the graph of a function f
[on the interval (a, b)] where the tangent line is parallel to the secant line that connects (a, f(a)) and (b, f(b)).
5. (a) See the I/D Test before Example 1 in Section 4.3.
(b) If the graph of f lies above all of its tangents on an interval I, then it is called concave upward on I.
(c) See the Concavity Test before Example 4 in Section 4.3.
(d) An inflection point is a point where a curve changes its direction of concavity. They can be found by determining the points
at which the second derivative changes sign.
6. (a) See the First Derivative Test after Example 1 in Section 4.3.
(b) See the Second Derivative Test before Example 6 in Section 4.3.
(c) See the note before Example 7 in Section 4.3.
7. (a) See Definitions 4.4.1 and 4.4.5.
(b) See Definitions 4.4.2 and 4.4.6.
(c) See Definition 4.4.7.
(d) See Definition 4.4.3.
8. Without calculus you could get misleading graphs that fail to show the most interesting features of a function.
See the discussion at the beginning of Section 4.5 and the first paragraph in Section 4.6.
172 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
9. (a) See Figure 2 in Section 4.8. (b) x2 = x1 − f(x1)
f 0(x1)(c) xn+1 = xn − f(xn)
f 0(xn)
(d) Newton’s method is likely to fail or to work very slowly when f 0(x1) is close to 0. It also fails when f 0(xi) is undefined,
such as with f(x) = 1/x− 2 and x1 = 1.
10. (a) See the definition at the beginning of Section 4.9.
(b) If F1 and F2 are both antiderivatives of f on an interval I, then they differ by a constant.
1. False. For example, take f(x) = x3, then f 0(x) = 3x2 and f 0(0) = 0, but f(0) = 0 is not a maximum or minimum;
(0, 0) is an inflection point.
3. False. For example, f(x) = x is continuous on (0, 1) but attains neither a maximum nor a minimum value on (0, 1).
Don’t confuse this with f being continuous on the closed interval [a, b], which would make the statement true.
5. True. This is an example of part (b) of the I/D Test.
7. False. f 0(x) = g0(x) ⇒ f(x) = g(x) +C. For example, if f(x) = x+2 and g(x) = x+1, then f 0(x) = g0(x) = 1,
but f(x) 6= g(x).
9. True. The graph of one such function is sketched.
11. True. Let x1 < x2 where x1, x2 ∈ I. Then f(x1) < f(x2) and g(x1) < g(x2) [since f and g are increasing on I ],
so (f + g)(x1) = f(x1) + g(x1) < f(x2) + g(x2) = (f + g)(x2).
13. False. Take f(x) = x and g(x) = x− 1. Then both f and g are increasing on (0, 1). But f(x) g(x) = x(x− 1) is not
increasing on (0, 1).
15. True. Let x1, x2 ∈ I and x1 < x2. Then f(x1) < f(x2) [f is increasing] ⇒ 1
f(x1)>
1
f(x2)[f is positive] ⇒
g(x1) > g(x2) ⇒ g(x) = 1/f(x) is decreasing on I.
17. True. If f is periodic, then there is a number p such that f(x+ p) = f(p) for all x. Differentiating gives
f 0(x) = f 0(x+ p) · (x+ p)0 = f 0(x+ p) · 1 = f 0(x+ p), so f 0 is periodic.
19. True. By the Mean Value Theorem, there exists a number c in (0, 1) such that f(1)− f(0) = f 0(c)(1− 0) = f 0(c).
Since f 0(c) is nonzero, f(1)− f(0) 6= 0, so f(1) 6= f(0).
CHAPTER 4 REVIEW ¤ 173
1. f(x) = x3 − 6x2 + 9x+ 1, [2, 4]. f 0(x) = 3x2 − 12x+ 9 = 3(x2 − 4x+ 3) = 3(x− 1)(x− 3). f 0(x) = 0 ⇒x = 1 or x = 3, but 1 is not in the interval. f 0(x) > 0 for 3 < x < 4 and f 0(x) < 0 for 2 < x < 3, so f(3) = 1 is a local
minimum value. Checking the endpoints, we find f(2) = 3 and f(4) = 5. Thus, f(3) = 1 is the absolute minimum value and
f(4) = 5 is the absolute maximum value.
3. f(x) = 3x− 4x2 + 1
, [−2, 2]. f 0(x) =(x2 + 1)(3)− (3x− 4)(2x)
(x2 + 1)2=−(3x2 − 8x− 3)
(x2 + 1)2=−(3x+ 1)(x− 3)
(x2 + 1)2.
f 0(x) = 0 ⇒ x = − 13
or x = 3, but 3 is not in the interval. f 0(x) > 0 for − 13< x < 2 and f 0(x) < 0 for
−2 < x < − 13
, so f − 13= −5
10/9= − 9
2is a local minimum value. Checking the endpoints, we find f(−2) = −2 and
f(2) = 25
. Thus, f − 13= − 9
2is the absolute minimum value and f(2) = 2
5is the absolute maximum value.
5. f(x) = x+ sin 2x, [0, π]. f 0(x) = 1 + 2 cos 2x = 0 ⇔ cos 2x = − 12⇔ 2x = 2π
3or 4π
3⇔ x = π
3or 2π
3.
f 00(x) = −4 sin 2x, so f 00 π3= −4 sin 2π
3= −2√3 < 0 and f 00 2π
3= −4 sin 4π
3= 2
√3 > 0, so
f π3= π
3+√32≈ 1.91 is a local maximum value and f 2π
3= 2π
3−√32≈ 1.23 is a local minimum value. Also f(0) = 0
and f(π) = π, so f(0) = 0 is the absolute minimum value and f(π) = π is the absolute maximum value.
7. limx→∞
3x4 + x− 56x4 − 2x2 + 1 = lim
x→∞
3 +1
x3− 5
x4
6− 2
x2+1
x4
=3 + 0 + 0
6− 0 + 0 =1
2
9. limx→−∞
√4x2 + 1
3x− 1 = limx→−∞
√4x2 + 1/
√x2
(3x− 1)/√x2 = limx→−∞
4 + 1/x2
−3 + 1/x [since−x = |x| = √x2 for x < 0]
=2
−3 + 0 = −2
3
11. limx→∞
√4x2 + 3x− 2x = lim
x→∞
√4x2 + 3x− 2x
1·√4x2 + 3x+ 2x√4x2 + 3x+ 2x
= limx→∞
(4x2 + 3x)− 4x2√4x2 + 3x+ 2x
= limx→∞
3x√4x2 + 3x+ 2x
= limx→∞
3x/√x2√
4x2 + 3x+ 2x /√x2
= limx→∞
3
4 + 3/x+ 2[since x = |x| = √x2 for x > 0]
=3
2 + 2=3
4
13. f(0) = 0, f 0(−2) = f 0(1) = f 0(9) = 0, limx→∞
f(x) = 0, limx→6
f(x) = −∞,
f 0(x) < 0 on (−∞,−2), (1, 6), and (9,∞), f 0(x) > 0 on (−2, 1) and (6, 9),
f 00(x) > 0 on (−∞, 0) and (12,∞), f 00(x) < 0 on (0, 6) and (6, 12)
174 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
15. f is odd, f 0(x) < 0 for 0 < x < 2, f 0(x) > 0 for x > 2,
f 00(x) > 0 for 0 < x < 3, f 00(x) < 0 for x > 3, limx→∞ f(x) = −2
17. y = f(x) = 2− 2x− x3 A. D = R B. y-intercept: f(0) = 2.
The x-intercept (approximately 0.770917) can be found using Newton’s
Method. C. No symmetry D. No asymptote
E. f 0(x) = −2− 3x2 = −(3x2 + 2) < 0, so f is decreasing on R.
F. No extreme value G. f 00(x) = −6x < 0 on (0,∞) and f 00(x) > 0 on
(−∞, 0), so f is CD on (0,∞) and CU on (−∞, 0). There is an IP at (0, 2).
H.
19. y = f(x) = x4 − 3x3 + 3x2 − x = x(x− 1)3 A. D = R B. y-intercept: f(0) = 0; x-intercepts: f(x) = 0 ⇔x = 0 or x = 1 C. No symmetry D. f is a polynomial function and hence, it has no asymptote.
E. f 0(x) = 4x3 − 9x2 + 6x− 1. Since the sum of the coefficients is 0, 1 is a root of f 0, so
f 0(x) = (x− 1) 4x2 − 5x+ 1 = (x− 1)2(4x− 1). f 0(x) < 0 ⇒ x < 14
, so f is decreasing on −∞, 14
and f is increasing on 14,∞ . F. f 0(x) does not change sign at x = 1, so
there is not a local extremum there. f 14= − 27
256is a local minimum value.
G. f 00(x) = 12x2 − 18x+ 6 = 6(2x− 1)(x− 1). f 00(x) = 0 ⇔ x = 12
or 1. f 00(x) < 0 ⇔ 12 < x < 1 ⇒ f is CD on 1
2 , 1 and CU on
−∞, 12
and (1,∞). There are inflection points at 12,− 1
16and (1, 0).
H.
21. y = f(x) =1
x(x− 3)2 A. D = {x | x 6= 0, 3} = (−∞, 0) ∪ (0, 3) ∪ (3,∞) B. No intercepts. C. No symmetry.
D. limx→±∞
1
x(x− 3)2 = 0, so y = 0 is a HA. limx→0+
1
x(x− 3)2 =∞, limx→0−
1
x(x− 3)2 = −∞, limx→3
1
x(x− 3)2 =∞,
so x = 0 and x = 3 are VA. E. f 0(x) = − (x− 3)2 + 2x(x− 3)
x2(x− 3)4 =3(1− x)
x2(x− 3)3 ⇒ f 0(x) > 0 ⇔ 1 < x < 3,
so f is increasing on (1, 3) and decreasing on (−∞, 0), (0, 1), and (3,∞).
F. Local minimum value f(1) = 14G. f 00(x) = 6(2x2 − 4x+ 3)
x3(x− 3)4 .
Note that 2x2 − 4x+ 3 > 0 for all x since it has negative discriminant.
So f 00(x) > 0 ⇔ x > 0 ⇒ f is CU on (0, 3) and (3,∞) and
CD on (−∞, 0). No IP
H.
CHAPTER 4 REVIEW ¤ 175
23. y = f(x) =x2
x+ 8= x− 8 + 64
x+ 8A. D = {x | x 6= −8} B. Intercepts are 0 C. No symmetry
D. limx→∞
x2
x+ 8=∞, but f(x)− (x− 8) = 64
x+ 8→ 0 as x→∞, so y = x− 8 is a slant asymptote.
limx→−8+
x2
x+ 8=∞ and lim
x→−8−x2
x+ 8= −∞, so x = −8 is a VA. E. f 0(x) = 1− 64
(x+ 8)2=
x(x+ 16)
(x+ 8)2> 0 ⇔
x > 0 or x < −16, so f is increasing on (−∞,−16) and (0,∞) and
decreasing on (−16,−8) and (−8, 0) .F. Local maximum value f(−16) = −32, local minimum value f(0) = 0
G. f 00(x) = 128/(x+ 8)3 > 0 ⇔ x > −8, so f is CU on (−8,∞) and
CD on (−∞,−8). No IP
H.
25. y = f(x) = x√2 + x A. D = [−2,∞) B. y-intercept: f(0) = 0; x-intercepts: −2 and 0 C. No symmetry
D. No asymptote E. f 0(x) = x
2√2 + x
+√2 + x =
1
2√2 + x
[x+ 2(2 + x)] =3x+ 4
2√2 + x
= 0 when x = − 43
, so f is
decreasing on −2,− 43
and increasing on − 43,∞ . F. Local minimum value f − 4
3= − 4
323= − 4
√6
9≈ −1.09,
no local maximum
G. f 00(x) =2√2 + x · 3− (3x+ 4) 1√
2 + x4(2 + x)
=6(2 + x)− (3x+ 4)
4(2 + x)3/2
=3x+ 8
4(2 + x)3/2
f 00(x) > 0 for x > −2, so f is CU on (−2,∞). No IP
H.
27. y = f(x) = sin2 x− 2 cosx A. D = R B. y-intercept: f(0) = −2 C. f(−x) = f(x), so f is symmetric with respect
to the y-axis. f has period 2π. D. No asymptote E. y0 = 2 sinx cosx+ 2 sinx = 2 sinx (cosx+ 1). y0 = 0 ⇔sinx = 0 or cosx = −1 ⇔ x = nπ or x = (2n+ 1)π. y0 > 0 when sinx > 0, since cosx+ 1 ≥ 0 for all x.
Therefore, y0 > 0 [and so f is increasing] on (2nπ, (2n+ 1)π); y0 < 0 [and so f is decreasing] on ((2n− 1)π, 2nπ).F. Local maximum values are f((2n+ 1)π) = 2; local minimum values are f(2nπ) = −2.
G. y0 = sin 2x+ 2 sinx ⇒ y00 = 2cos 2x+ 2cosx = 2(2 cos2 x− 1) + 2 cosx = 4 cos2 x+ 2cosx− 2= 2(2 cos2 x+ cosx− 1) = 2(2 cosx− 1)(cosx+ 1)
y00 = 0 ⇔ cosx = 12
or −1 ⇔ x = 2nπ ± π3
or x = (2n+ 1)π.
y00 > 0 [and so f is CU] on 2nπ − π3, 2nπ + π
3; y00 ≤ 0 [and so f is CD]
on 2nπ + π3, 2nπ + 5π
3. There are inflection points at 2nπ ± π
3,− 1
4.
H.
176 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
29. f(x) = x2 − 1x3
⇒ f 0(x) =x3(2x)− x2 − 1 3x2
x6=3− x2
x4⇒
f 00(x) =x4(−2x)− 3− x2 4x3
x8=2x2 − 12
x5
Estimates: From the graphs of f 0 and f 00, it appears that f is increasing on
(−1.73, 0) and (0, 1.73) and decreasing on (−∞,−1.73) and (1.73,∞);f has a local maximum of about f(1.73) = 0.38 and a local minimum of about
f(−1.7) = −0.38; f is CU on (−2.45, 0) and (2.45,∞), and CD on
(−∞,−2.45) and (0, 2.45); and f has inflection points at about
(−2.45,−0.34) and (2.45, 0.34).
Exact: Now f 0(x) =3− x2
x4is positive for 0 < x2 < 3, that is, f is increasing
on −√3, 0 and 0,√3 ; and f 0 (x) is negative (and so f is decreasing) on
−∞,−√3 and√3,∞ . f 0 (x) = 0 when x = ±√3.
f 0 goes from positive to negative at x =√3, so f has a local maximum of
f√3 =
(√3 )2− 1(√3 )3
= 2√3
9 ; and since f is odd, we know that maxima on the
interval (0,∞) correspond to minima on (−∞, 0), so f has a local minimum of
f −√3 = − 2√3
9. Also, f 00 (x) = 2x2 − 12
x5is positive (so f is CU) on
−√6, 0 and√6,∞ , and negative (so f is CD) on −∞,−√6 and
0,√6 . There are IP at
√6, 5
√6
36and −√6,− 5
√6
36.
31. f(x) = 3x6 − 5x5 + x4 − 5x3 − 2x2 + 2 ⇒ f 0(x) = 18x5 − 25x4 + 4x3 − 15x2 − 4x ⇒
f 00(x) = 90x4 − 100x3 + 12x2 − 30x− 4
From the graphs of f 0 and f 00, it appears that f is increasing on (−0.23, 0) and (1.62,∞) and decreasing on (−∞,−0.23)
and (0, 1.62); f has a local maximum of about f(0) = 2 and local minima of about f(−0.23) = 1.96 and f(1.62) = −19.2;
CHAPTER 4 REVIEW ¤ 177
f is CU on (−∞,−0.12) and (1.24,∞) and CD on (−0.12, 1.24); and f has inflection points at about (−0.12, 1.98) and
(1.24,−12.1).
33. Let f(x) = 3x+ 2cosx+ 5. Then f(0) = 7 > 0 and f(−π) = −3π − 2 + 5 = −3π + 3 = −3(π − 1) < 0, and since f is
continuous on R (hence on [−π, 0]), the Intermediate Value Theorem assures us that there is at least one zero of f in [−π, 0].Now f 0(x) = 3− 2 sinx > 0 implies that f is increasing on R, so there is exactly one zero of f , and hence, exactly one real
root of the equation 3x+ 2cosx+ 5 = 0.
35. Since f is continuous on [32, 33] and differentiable on (32, 33), then by the Mean Value Theorem there exists a number c in
(32, 33) such that f 0(c) = 15c−4/5 =
5√33− 5
√32
33− 32 = 5√33− 2, but 1
5c−4/5 > 0 ⇒ 5
√33− 2 > 0 ⇒ 5
√33 > 2. Also
f 0 is decreasing, so that f 0(c) < f 0(32) = 15(32)−4/5 = 0.0125 ⇒ 0.0125 > f 0(c) = 5
√33− 2 ⇒ 5
√33 < 2.0125.
Therefore, 2 < 5√33 < 2.0125.
37. (a) g(x) = f(x2) ⇒ g0(x) = 2xf 0(x2) by the Chain Rule. Since f 0(x) > 0 for all x 6= 0, we must have f 0(x2) > 0 for
x 6= 0, so g0(x) = 0 ⇔ x = 0. Now g0(x) changes sign (from negative to positive) at x = 0, since one of its factors,
f 0(x2), is positive for all x, and its other factor, 2x, changes from negative to positive at this point, so by the First
Derivative Test, f has a local and absolute minimum at x = 0.
(b) g0(x) = 2xf 0(x2) ⇒ g00(x) = 2[xf 00(x2)(2x) + f 0(x2)] = 4x2f 00(x2) + 2f 0(x2) by the Product Rule and the Chain
Rule. But x2 > 0 for all x 6= 0, f 00(x2) > 0 [since f is CU for x > 0], and f 0(x2) > 0 for all x 6= 0, so since all of its
factors are positive, g00(x) > 0 for x 6= 0. Whether g00(0) is positive or 0 doesn’t matter [since the sign of g00 does not
change there]; g is concave upward on R.
39. If B = 0, the line is vertical and the distance from x = −C
Ato (x1, y1) is x1 +
C
A=|Ax1 +By1 +C|√
A2 +B2, so assume
B 6= 0. The square of the distance from (x1, y1) to the line is f(x) = (x− x1)2 + (y − y1)
2 where Ax+By + C = 0, so
we minimize f(x) = (x− x1)2 + −A
Bx− C
B− y1
2
⇒ f 0(x) = 2 (x− x1) + 2 −A
Bx− C
B− y1 −A
B.
f 0(x) = 0 ⇒ x =B2x1 −ABy1 −AC
A2 +B2and this gives a minimum since f 00(x) = 2 1 +
A2
B2> 0. Substituting
this value of x into f(x) and simplifying gives f(x) = (Ax1 +By1 + C)2
A2 +B2, so the minimum distance is
f(x) =|Ax1 +By1 + C|√
A2 +B2.
178 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
41. By similar triangles, yx=
r√x2 − 2rx , so the area of the triangle is
A(x) = 12(2y)x = xy =
rx2√x2 − 2rx ⇒
A0(x) =2rx
√x2 − 2rx− rx2(x− r)/
√x2 − 2rx
x2 − 2rx =rx2 (x− 3r)(x2 − 2rx)3/2
= 0
when x = 3r.
A0(x) < 0 when 2r < x < 3r, A0(x) > 0 when x > 3r. So x = 3r gives a minimum and A(3r) = r(9r2)√3 r
= 3√3 r2.
43. We minimize L(x) = |PA|+ |PB|+ |PC| = 2√x2 + 16 + (5− x),
0 ≤ x ≤ 5. L0(x) = 2x√x2 + 16 − 1 = 0 ⇔ 2x =
√x2 + 16 ⇔
4x2 = x2 + 16 ⇔ x = 4√3
. L(0) = 13, L 4√3≈ 11.9, L(5) ≈ 12.8, so the
minimum occurs when x = 4√3≈ 2.3.
45. v = KL
C+
C
L⇒ dv
dL=
K
2 (L/C) + (C/L)
1
C− C
L2= 0 ⇔ 1
C=
C
L2⇔ L2 = C2 ⇔ L = C.
This gives the minimum velocity since v0 < 0 for 0 < L < C and v0 > 0 for L > C.
47. Let x denote the number of $1 decreases in ticket price. Then the ticket price is $12− $1(x), and the average attendance is
11,000 + 1000(x). Now the revenue per game is
R(x) = (price per person)× (number of people per game)
= (12− x)(11,000 + 1000x) = −1000x2 + 1000x+ 132,000
for 0 ≤ x ≤ 4 [since the seating capacity is 15,000] ⇒ R0(x) = −2000x+ 1000 = 0 ⇔ x = 0.5. This is a
maximum since R00(x) = −2000 < 0 for all x. Now we must check the value of R(x) = (12− x)(11,000 + 1000x) at
x = 0.5 and at the endpoints of the domain to see which value of x gives the maximum value of R.
R(0) = (12)(11,000) = 132,000, R(0.5) = (11.5)(11,500) = 132,250, and R(4) = (8)(15,000) = 120,000. Thus, the
maximum revenue of $132,250 per game occurs when the average attendance is 11,500 and the ticket price is $11.50.
49. f(x) = x5 − x4 + 3x2 − 3x− 2 ⇒ f 0(x) = 5x4 − 4x3 + 6x− 3, so xn+1 = xn − x5n − x4n + 3x2n − 3xn − 2
5x4n − 4x3n + 6xn − 3 .
Now x1 = 1 ⇒ x2 = 1.5 ⇒ x3 ≈ 1.343860 ⇒ x4 ≈ 1.300320 ⇒ x5 ≈ 1.297396 ⇒
x6 ≈ 1.297383 ≈ x7, so the root in [1, 2] is 1.297383, to six decimal places.
CHAPTER 4 REVIEW ¤ 179
51. f(t) = cos t+ t− t2 ⇒ f 0(t) = − sin t+ 1− 2t. f 0(t) exists for all
t, so to find the maximum of f , we can examine the zeros of f 0.
From the graph of f 0, we see that a good choice for t1 is t1 = 0.3.
Use g(t) = − sin t+ 1− 2t and g0(t) = − cos t− 2 to obtain
t2 ≈ 0.33535293, t3 ≈ 0.33541803 ≈ t4. Since f 00(t) = − cos t− 2 < 0
for all t, f(0.33541803) ≈ 1.16718557 is the absolute maximum.
53. f 0(x) =√x3 +
3√x2 = x3/2 + x2/3 ⇒ f(x) =
x5/2
5/2+
x5/3
5/3+C = 2
5x5/2 + 3
5x5/3 + C
55. f 0(t) = 2t− 3 sin t ⇒ f(t) = t2 + 3cos t+C.
f(0) = 3 +C and f(0) = 5 ⇒ C = 2, so f(t) = t2 + 3cos t+ 2.
57. f 00(x) = 1− 6x+ 48x2 ⇒ f 0(x) = x− 3x2 + 16x3 + C. f 0(0) = C and f 0(0) = 2 ⇒ C = 2, so
f 0(x) = x− 3x2 + 16x3 + 2 and hence, f(x) = 12x
2 − x3 + 4x4 + 2x+D.
f(0) = D and f(0) = 1 ⇒ D = 1, so f(x) = 12x2 − x3 + 4x4 + 2x+ 1.
59. v(t) = s0(t) = 2t− sin t ⇒ s(t) = t2 + cos t+ C.
s(0) = 0 + 1 + C = C + 1 and s(0) = 3 ⇒ C + 1 = 3 ⇒ C = 2, so s(t) = t2 + cos t+ 2.
61. f(x) = x2 sin(x2), 0 ≤ x ≤ π
63. Choosing the positive direction to be upward, we have a(t) = −9.8 ⇒ v(t) = −9.8t+ v0, but v(0) = 0 = v0 ⇒v(t) = −9.8t = s0(t) ⇒ s(t) = −4.9t2 + s0, but s(0) = s0 = 500 ⇒ s(t) = −4.9t2 + 500. When s = 0,
−4.9t2 + 500 = 0 ⇒ t1 =5004.9 ≈ 10.1 ⇒ v(t1) = −9.8 500
4.9 ≈ −98.995 m/s. Since the canister has been
designed to withstand an impact velocity of 100 m/s, the canister will not burst.
65. (a) The cross-sectional area of the rectangular beam is
A = 2x · 2y = 4xy = 4x√100− x2, 0 ≤ x ≤ 10, so
dA
dx= 4x 1
2(100− x2)−1/2(−2x) + (100− x2)1/2 · 4
=−4x2
(100− x2)1/2+ 4(100− x2)1/2 =
4[−x2 + 100− x2 ]
(100− x2)1/2.
dA
dx= 0 when −x2 + 100− x2 = 0 ⇒ x2 = 50 ⇒ x =
√50 ≈ 7.07 ⇒ y = 100− √
502=√50.
Since A(0) = A(10) = 0, the rectangle of maximum area is a square.
180 ¤ CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
(b) The cross-sectional area of each rectangular plank (shaded in the figure) is
A = 2x y −√50 = 2x√100− x2 −√50 , 0 ≤ x ≤ √50, so
dA
dx= 2
√100− x2 −√50 + 2x 1
2(100− x2)−1/2(−2x)
= 2(100− x2)1/2 − 2√50− 2x2
(100− x2)1/2
Set dAdx
= 0: (100− x2)−√50 (100− x2)1/2 − x2 = 0 ⇒ 100− 2x2 = √50 (100− x2)1/2 ⇒
10,000− 400x2 + 4x4 = 50(100− x2) ⇒ 4x4 − 350x2 + 5000 = 0 ⇒ 2x4 − 175x2 + 2500 = 0 ⇒
x2 =175±√10,625
4≈ 69.52 or 17.98 ⇒ x ≈ 8.34 or 4.24. But 8.34 >
√50, so x1 ≈ 4.24 ⇒
y −√50 = 100− x21 −√50 ≈ 1.99. Each plank should have dimensions about 8 12 inches by 2 inches.
(c) From the figure in part (a), the width is 2x and the depth is 2y, so the strength is
S = k(2x)(2y)2 = 8kxy2 = 8kx(100− x2) = 800kx− 8kx3, 0 ≤ x ≤ 10. dS/dx = 800k − 24kx2 = 0 when
24kx2 = 800k ⇒ x2 = 1003
⇒ x = 10√3⇒ y = 200
3= 10
√2√3=√2x. Since S(0) = S(10) = 0, the
maximum strength occurs when x = 10√3
. The dimensions should be 20√3≈ 11.55 inches by 20
√2√3≈ 16.33 inches.
PROBLEMS PLUS
1. Let f(x) = sinx− cosx on [0, 2π] since f has period 2π. f 0(x) = cosx+ sinx = 0 ⇔ cosx = − sinx ⇔
tanx = −1 ⇔ x = 3π4
or 7π4
. Evaluating f at its critical numbers and endpoints, we get f(0) = −1, f 3π4
=√2,
f 7π4
= −√2, and f(2π) = −1. So f has absolute maximum value√2 and absolute minimum value −√2. Thus,
−√2 ≤ sinx− cosx ≤ √2 ⇒ |sinx− cosx| ≤ √2.
3. First we show that x(1− x) ≤ 14 for all x. Let f(x) = x(1− x) = x− x2. Then f 0(x) = 1− 2x. This is 0 when x = 1
2 and
f 0(x) > 0 for x < 12
, f 0(x) < 0 for x > 12
, so the absolute maximum of f is f 12= 1
4. Thus, x(1− x) ≤ 1
4for all x.
Now suppose that the given assertion is false, that is, a(1− b) > 14
and b(1− a) > 14
. Multiply these inequalities:
a(1− b)b(1− a) > 116
⇒ [a(1− a)][b(1− b)] > 116
. But we know that a(1− a) ≤ 14
and b(1− b) ≤ 14⇒
[a(1− a)][b(1− b)] ≤ 116 . Thus, we have a contradiction, so the given assertion is proved.
5. Differentiating x2 + xy + y2 = 12 implicitly with respect to x gives 2x+ y + xdy
dx+ 2y
dy
dx= 0, so dy
dx= −2x+ y
x+ 2y.
At a highest or lowest point, dydx
= 0 ⇔ y = −2x. Substituting −2x for y in the original equation gives
x2 + x(−2x) + (−2x)2 = 12, so 3x2 = 12 and x = ±2. If x = 2, then y = −2x = −4, and if x = −2 then y = 4.
Thus, the highest and lowest points are (−2, 4) and (2,−4).
7. f(x) = 1
1 + |x| +1
1 + |x− 2|
=
⎧⎪⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎪⎩
1
1− x+
1
1− (x− 2) if x < 0
1
1 + x+
1
1− (x− 2) if 0 ≤ x < 2
1
1 + x+
1
1 + (x− 2) if x ≥ 2
⇒ f 0 (x) =
⎧⎪⎪⎪⎪⎪⎪⎪⎨⎪⎪⎪⎪⎪⎪⎪⎩
1
(1− x)2+
1
(3− x)2if x < 0
−1(1 + x)2
+1
(3− x)2if 0 < x < 2
−1(1 + x)2
− 1
(x− 1)2 if x > 2
We see that f 0(x) > 0 for x < 0 and f 0(x) < 0 for x > 2. For 0 < x < 2, we have
f 0(x) =1
(3− x)2− 1
(x+ 1)2=(x2 + 2x+ 1)− (x2 − 6x+ 9)
(3− x)2(x+ 1)2=
8 (x− 1)(3− x)2(x+ 1)2
, so f 0(x) < 0 for 0 < x < 1,
f 0(1) = 0 and f 0(x) > 0 for 1 < x < 2. We have shown that f 0(x) > 0 for x < 0; f 0(x) < 0 for 0 < x < 1; f 0(x) > 0 for
1 < x < 2; and f 0(x) < 0 for x > 2. Therefore, by the First Derivative Test, the local maxima of f are at x = 0 and x = 2,
where f takes the value 43
. Therefore, 43
is the absolute maximum value of f .
9. A = x1, x21 and B = x2, x
22 , where x1 and x2 are the solutions of the quadratic equation x2 = mx+ b. Let P = x, x2
and set A1 = (x1, 0), B1 = (x2, 0), and P1 = (x, 0). Let f(x) denote the area of triangle PAB. Then f(x) can be expressed
181
182 ¤ CHAPTER 4 PROBLEMS PLUS
in terms of the areas of three trapezoids as follows:
f(x) = area (A1ABB1)− area (A1APP1)− area (B1BPP1)
= 12x21 + x22 (x2 − x1)− 1
2x21 + x2 (x− x1)− 1
2x2 + x22 (x2 − x)
After expanding and canceling terms, we get
f(x) = 12x2x
21 − x1x
22 − xx21 + x1x
2 − x2x2 + xx22 = 1
2x21(x2 − x) + x22(x− x1) + x2(x1 − x2)
f 0(x) = 12−x21 + x22 + 2x(x1 − x2) . f 00(x) = 1
2[2(x1 − x2)] = x1 − x2 < 0 since x2 > x1.
f 0(x) = 0 ⇒ 2x(x1 − x2) = x21 − x22 ⇒ xP =12 (x1 + x2).
f(xP ) =12x21
12(x2 − x1) + x22
12(x2 − x1) +
14(x1 + x2)
2(x1 − x2)
= 12
12(x2 − x1) x
21 + x22 − 1
4(x2 − x1)(x1 + x2)
2 = 18(x2 − x1) 2 x21 + x22 − x21 + 2x1x2 + x22
= 18 (x2 − x1) x
21 − 2x1x2 + x22 = 1
8 (x2 − x1)(x1 − x2)2 = 1
8 (x2 − x1)(x2 − x1)2 = 1
8 (x2 − x1)3
To put this in terms of m and b, we solve the system y = x21 and y = mx1 + b, giving us x21 −mx1 − b = 0 ⇒
x1 =12m−√m2 + 4b . Similarly, x2 = 1
2m+
√m2 + 4b . The area is then 1
8 (x2 − x1)3 = 1
8
√m2 + 4b
3,
and is attained at the point P xP , x2P = P 1
2m, 1
4m2 .
Note: Another way to get an expression for f(x) is to use the formula for an area of a triangle in terms of the coordinates of
the vertices: f(x) = 12
x2x21 − x1x
22 + x1x
2 − xx21 + xx22 − x2x2 .
11. f(x) = a2 + a− 6 cos 2x+ (a− 2)x+ cos 1 ⇒ f 0(x) = − a2 + a− 6 sin 2x (2) + (a− 2). The derivative exists
for all x, so the only possible critical points will occur where f 0(x) = 0 ⇔ 2(a− 2)(a+ 3) sin 2x = a− 2 ⇔
either a = 2 or 2(a+ 3) sin 2x = 1, with the latter implying that sin 2x = 1
2(a+ 3). Since the range of sin 2x is [−1, 1],
this equation has no solution whenever either 1
2(a+ 3)< −1 or 1
2(a+ 3)> 1. Solving these inequalities, we get
− 72< a < − 5
2.
13. (a) Let y = |AD|, x = |AB|, and 1/x = |AC|, so that |AB| · |AC| = 1. We compute
the areaA of4ABC in two ways.
First, A = 12 |AB| |AC| sin 2π
3 = 12 · 1 ·
√32 =
√34 .
Second,
A = (area of4ABD) + (area of4ACD) = 12|AB| |AD| sin π
3+ 1
2|AD| |AC| sin π
3
= 12xy
√32+ 1
2y(1/x)
√32=√34y(x+ 1/x)
Equating the two expressions for the area, we get√34 y x+
1
x=√34 ⇔ y =
1
x+ 1/x=
x
x2 + 1, x > 0.
Another method: Use the Law of Sines on the triangles ABD and ABC. In4ABD, we have ∠A+∠B + ∠D = 180◦
⇔ 60◦ + α+ ∠D = 180◦ ⇔ ∠D = 120◦ − α. Thus,
CHAPTER 4 PROBLEMS PLUS ¤ 183
x
y=sin(120◦ − α)
sinα=sin 120◦ cosα− cos 120◦ sinα
sinα=
√32cosα+ 1
2sinα
sinα⇒ x
y=√32 cotα+
12 , and by a
similar argument with4ABC,√32cotα = x2 + 1
2. Eliminating cotα gives x
y= x2 + 1
2+ 1
2⇒
y =x
x2 + 1, x > 0.
(b) We differentiate our expression for y with respect to x to find the maximum:
dy
dx=(x2 + 1)− x(2x)
(x2 + 1)2=
1− x2
(x2 + 1)2= 0 when x = 1. This indicates a maximum by the First Derivative Test, since
y0(x) > 0 for 0 < x < 1 and y0(x) < 0 for x > 1, so the maximum value of y is y(1) = 12 .
15. (a) A = 12bh with sin θ = h/c, so A = 1
2bc sin θ. But A is a constant,
so differentiating this equation with respect to t, we get
dA
dt= 0 =
1
2bc cos θ
dθ
dt+ b
dc
dtsin θ +
db
dtc sin θ ⇒
bc cos θdθ
dt= − sin θ b
dc
dt+ c
db
dt⇒ dθ
dt= − tan θ 1
c
dc
dt+1
b
db
dt.
(b) We use the Law of Cosines to get the length of side a in terms of those of b and c, and then we differentiate implicitly with
respect to t: a2 = b2 + c2 − 2bc cos θ ⇒
2ada
dt= 2b
db
dt+ 2c
dc
dt− 2 bc(− sin θ) dθ
dt+ b
dc
dtcos θ +
db
dtc cos θ ⇒
da
dt=1
abdb
dt+ c
dc
dt+ bc sin θ
dθ
dt− b
dc
dtcos θ − c
db
dtcos θ . Now we substitute our value of a from the Law of
Cosines and the value of dθ/dt from part (a), and simplify (primes signify differentiation by t):
da
dt=
bb0 + cc0 + bc sin θ [− tan θ(c0/c+ b0/b)]− (bc0 + cb0)(cos θ)√b2 + c2 − 2bc cos θ
=bb0 + cc0 − [sin2 θ (bc0 + cb0) + cos2 θ (bc0 + cb0)]/ cos θ√
b2 + c2 − 2bc cos θ =bb0 + cc0 − (bc0 + cb0) sec θ√
b2 + c2 − 2bc cos θ
17. (a) Distance = rate× time, so time = distance/rate. T1 =D
c1, T2 =
2 |PR|c1
+|RS|c2
=2h sec θ
c1+
D − 2h tan θc2
,
T3 =2 h2 + D2/4
c1=
√4h2 +D2
c1.
(b) dT2dθ
=2h
c1· sec θ tan θ − 2h
c2sec2 θ = 0 when 2h sec θ 1
c1tan θ − 1
c2sec θ = 0 ⇒
1
c1
sin θ
cos θ− 1
c2
1
cos θ= 0 ⇒ sin θ
c1 cos θ=
1
c2 cos θ⇒ sin θ =
c1c2
. The First Derivative Test shows that this gives
a minimum.
(c) Using part (a) with D = 1 and T1 = 0.26, we have T1 =D
c1⇒ c1 =
10.26
≈ 3.85 km/s. T3 =√4h2 +D2
c1⇒
4h2 +D2 = T 23 c21 ⇒ h = 1
2T 23c21 −D2 = 1
2(0.34)2(1/0.26)2 − 12 ≈ 0.42 km. To find c2, we use sin θ = c1
c2
184 ¤ CHAPTER 4 PROBLEMS PLUS
from part (b) and T2 =2h sec θ
c1+
D − 2h tan θc2
from part (a). From the figure,
sin θ =c1c2
⇒ sec θ =c2
c22 − c21and tan θ = c1
c22 − c21, so
T2 =2hc2
c1 c22 − c21+
D c22 − c21 − 2hc1c2 c22 − c21
. Using the values for T2 [given as 0.32],
h, c1, and D, we can graph Y1 = T2 and Y2 =2hc2
c1 c22 − c21+
D c22 − c21 − 2hc1c2 c22 − c21
and find their intersection points.
Doing so gives us c2 ≈ 4.10 and 7.66, but if c2 = 4.10, then θ = arcsin(c1/c2) ≈ 69.6◦, which implies that point S is to
the left of point R in the diagram. So c2 = 7.66 km/s.
19. Let a = |EF | and b = |BF | as shown in the figure. Since = |BF |+ |FD|,|FD| = − b. Now
|ED|= |EF |+ |FD| = a+ − b =√r2 − x2 + − (d− x)2 + a2
=√r2 − x2 + − (d− x)2 +
√r2 − x2
2
=√r2 − x2 + −√d2 − 2dx+ x2 + r2 − x2
Let f(x) =√r2 − x2 + −√d2 + r2 − 2dx.
f 0(x) = 12 (r
2 − x2)−1/2(−2x)− 12(d
2 + r2 − 2dx)−1/2(−2d) = −x√r2 − x2
+d√
d2 + r2 − 2dx .
f 0(x) = 0 ⇒ x√r2 − x2
=d√
d2 + r2 − 2dx ⇒ x2
r2 − x2=
d2
d2 + r2 − 2dx ⇒
d2x2 + r2x2 − 2dx3 = d2r2 − d2x2 ⇒ 0 = 2dx3 − 2d2x2 − r2x2 + d2r2 ⇒0 = 2dx2(x− d)− r2(x2 − d2) ⇒ 0 = 2dx2(x− d)− r2(x+ d)(x− d) ⇒ 0 = (x− d)[2dx2 − r2(x+ d)]
But d > r > x, so x 6= d. Thus, we solve 2dx2 − r2x − dr2 = 0 for x:
x =−(−r2)± (−r2)2 − 4(2d)(−dr2)
2(2d)=
r2 ±√r4 + 8d2r24d
. Because√r4 + 8d2r2 > r2, the “negative” can be
discarded. Thus, x = r2 +√r2√r2 + 8d2
4d=
r2 + r√r2 + 8d2
4d[r > 0] =
r
4dr +
√r2 + 8d2 . The maximum value
of |ED| occurs at this value of x.
21. V = 43πr3 ⇒ dV
dt= 4πr2
dr
dt. But dV
dtis proportional to the surface area, so dV
dt= k · 4πr2 for some constant k.
Therefore, 4πr2 drdt= k · 4πr2 ⇔ dr
dt= k = constant. An antiderivative of k with respect to t is kt, so r = kt+ C.
When t = 0, the radius r must equal the original radius r0, so C = r0, and r = kt+ r0. To find k we use the fact that when
t = 3, r = 3k + r0 and V = 12V0 ⇒ 4
3π(3k + r0)
3 = 12· 43πr30 ⇒ (3k + r0)
3 = 12r30 ⇒ 3k + r0 =
13√2r0 ⇒
k = 13r0
13√2− 1 . Since r = kt+ r0, r = 1
3r0
13√2− 1 t+ r0. When the snowball has melted completely we have
r = 0 ⇒ 13r0
13√2− 1 t+ r0 = 0 which gives t = 3 3
√2
3√2− 1 . Hence, it takes 3 3
√2
3√2− 1 − 3 =
33√2− 1 ≈ 11 h 33 min
longer.